I have a H8DM8-2 motherboard with a pair of AOC-SAT2-MV8 SATA controller cards in a 16-disk Supermicro chassis. I''m running OpenSolaris 2008.11, and the machine performs very well unless I start to copy a large amount of data to the ZFS (software raid) array that''s on the Supermicro SATA controllers. If I do this, the machine inevitably reboots. What can I do to troubleshoot? The BIOS of the motherboard and the SATA card firmware are fully updated. I''m running the latest stable OpenSolaris, and see nothing amiss in the system logs when this happens. I''ve enabled savecore and debug-level syslog, but am getting no indicators from Solaris as to what''s wrong. Interestingly, I can push the same amount of data to the mirror boot disks, which are on the board''s built-in nVidia SATA controller without issue. The vdev I''m pushing to is a 5-disk raidz2 with 2 hot spares. Help! :)
The copy operation will make all the disks start seeking at the same time and will make your CPU activity jump to a significant percentage to compute the ZFS checksum and RAIDZ parity. I think you could be overloading your PSU because of the sudden increase in power consumption... However if you are *not* using SATA staggered spin-up, then the above theory is unlikely because spinning up consumes much more power than when seeking. So, in a sense, a successful boot proves your PSU is powerful enough. Trying reproducing the problem by copying data on a smaller number of disks. You tried 2 and 16. Try 8. -marc
I''m working on testing this some more by doing a savecore -L right after I start the copy. BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis supports 16, but isn''t fully populated). So far as I know, there is no spinup happening - these are not RAID controllers, just dumb SATA JBOD controllers, so I don''t think they control drive spin in any particular way. Correct me if I''m wrong, of course. On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote:> The copy operation will make all the disks start seeking at the same time and > will make your CPU activity jump to a significant percentage to compute the > ZFS checksum and RAIDZ parity. I think you could be overloading your PSU > because of the sudden increase in power consumption... > > However if you are *not* using SATA staggered spin-up, then the above theory > is unlikely because spinning up consumes much more power than when seeking. > So, in a sense, a successful boot proves your PSU is powerful enough. > > Trying reproducing the problem by copying data on a smaller number of disks. > You tried 2 and 16. Try 8. > > -marc > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >
I''m attaching a screenshot of the console just before reboot. The dump doesn''t seem to be working, or savecore isn''t working. On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote:> I''m working on testing this some more by doing a savecore -L right > after I start the copy. > > BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis > supports 16, but isn''t fully populated). > > So far as I know, there is no spinup happening - these are not RAID > controllers, just dumb SATA JBOD controllers, so I don''t think they > control drive spin in any particular way. ?Correct me if I''m wrong, of > course. > > > > On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >> The copy operation will make all the disks start seeking at the same time and >> will make your CPU activity jump to a significant percentage to compute the >> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >> because of the sudden increase in power consumption... >> >> However if you are *not* using SATA staggered spin-up, then the above theory >> is unlikely because spinning up consumes much more power than when seeking. >> So, in a sense, a successful boot proves your PSU is powerful enough. >> >> Trying reproducing the problem by copying data on a smaller number of disks. >> You tried 2 and 16. Try 8. >> >> -marc >> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> >-------------- next part -------------- A non-text attachment was scrubbed... Name: IMG_0146.JPG Type: image/jpeg Size: 647193 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090311/1c6bbe11/attachment.jpe>
Remco Lengers
2009-Mar-11 19:08 UTC
[zfs-discuss] reboot when copying large amounts of data
Something is not right in the IO space. The messages talk about vendor ID = 11AB 0x11AB Marvell Semiconductor TMC Research Vendor Id: 0x1030 Short Name: TMC Does "fmdump -eV" give any clue when the box comes back up? ..Remco Blake wrote:> I''m attaching a screenshot of the console just before reboot. The > dump doesn''t seem to be working, or savecore isn''t working. > > On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >> I''m working on testing this some more by doing a savecore -L right >> after I start the copy. >> >> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >> supports 16, but isn''t fully populated). >> >> So far as I know, there is no spinup happening - these are not RAID >> controllers, just dumb SATA JBOD controllers, so I don''t think they >> control drive spin in any particular way. Correct me if I''m wrong, of >> course. >> >> >> >> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >>> The copy operation will make all the disks start seeking at the same time and >>> will make your CPU activity jump to a significant percentage to compute the >>> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >>> because of the sudden increase in power consumption... >>> >>> However if you are *not* using SATA staggered spin-up, then the above theory >>> is unlikely because spinning up consumes much more power than when seeking. >>> So, in a sense, a successful boot proves your PSU is powerful enough. >>> >>> Trying reproducing the problem by copying data on a smaller number of disks. >>> You tried 2 and 16. Try 8. >>> >>> -marc >>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>> >> >> ------------------------------------------------------------------------ >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
fmdump is not helping much: root at host:~# fmdump -eV TIME CLASS fmdump: /var/fm/fmd/errlog is empty comparing that screenshot to the output of cfgadm is interesting - looks like the controller(s): root at host:~# cfgadm -v Ap_Id Receptacle Occupant Condition Information When Type Busy Phys_Id sata4/0::dsk/c4t0d0 connected configured ok Mod: ST3250310NS FRev: SN06 SN: 9SF06CZZ unavailable disk n /devices/pci at 0,0/pci15d9,1611 at 5:0 sata4/1::dsk/c4t1d0 connected configured ok Mod: ST3250310NS FRev: SN06 SN: 9SF06BC8 unavailable disk n /devices/pci at 0,0/pci15d9,1611 at 5:1 sata5/0 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci15d9,1611 at 5,1:0 sata5/1::dsk/c7t1d0 connected configured ok Mod: WDC WD10EACS-00D6B0 FRev: 01.01A01 SN: WD-WCAU40244615 unavailable disk n /devices/pci at 0,0/pci15d9,1611 at 5,1:1 sata6/0 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci15d9,1611 at 5,2:0 sata6/1 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci15d9,1611 at 5,2:1 sata7/0 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:0 sata7/1 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:1 sata7/2::dsk/c5t2d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0376631 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:2 sata7/3::dsk/c5t3d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0350798 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:3 sata7/4::dsk/c5t4d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0403574 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:4 sata7/5::dsk/c5t5d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0312592 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:5 sata7/6::dsk/c5t6d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0399779 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:6 sata7/7::dsk/c5t7d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT0441660 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 4:7 sata8/0::dsk/c6t0d0 connected configured ok Mod: WDC WD7500AYYS-01RCA0 FRev: 30.04G30 SN: WD-WCAPT1000344 unavailable disk n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:0 sata8/1 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:1 sata8/2 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:2 sata8/3 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:3 sata8/4 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:4 sata8/5 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:5 sata8/6 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:6 sata8/7 empty unconfigured ok unavailable sata-port n /devices/pci at 0,0/pci10de,376 at a/pci1033,125 at 0/pci11ab,11ab at 6:7 On Wed, Mar 11, 2009 at 2:40 PM, Blake <blake.irvin at gmail.com> wrote:> I''m attaching a screenshot of the console just before reboot. ?The > dump doesn''t seem to be working, or savecore isn''t working. > > On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >> I''m working on testing this some more by doing a savecore -L right >> after I start the copy. >> >> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >> supports 16, but isn''t fully populated). >> >> So far as I know, there is no spinup happening - these are not RAID >> controllers, just dumb SATA JBOD controllers, so I don''t think they >> control drive spin in any particular way. ?Correct me if I''m wrong, of >> course. >> >> >> >> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >>> The copy operation will make all the disks start seeking at the same time and >>> will make your CPU activity jump to a significant percentage to compute the >>> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >>> because of the sudden increase in power consumption... >>> >>> However if you are *not* using SATA staggered spin-up, then the above theory >>> is unlikely because spinning up consumes much more power than when seeking. >>> So, in a sense, a successful boot proves your PSU is powerful enough. >>> >>> Trying reproducing the problem by copying data on a smaller number of disks. >>> You tried 2 and 16. Try 8. >>> >>> -marc >>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>> >> >
I think that TMC Research is the company that designed the Supermicro-branded controller card that has the Marvell SATA controller chip on it. Googling around I see connections between Supermicro and TMC. This is the card: http://www.supermicro.com/products/accessories/addon/AOC-SAT2-MV8.cfm On Wed, Mar 11, 2009 at 3:08 PM, Remco Lengers <remco at lengers.com> wrote:> Something is not right in the IO space. The messages talk about > > vendor ID = 11AB > > 0x11AB ?Marvell Semiconductor > > TMC Research > > Vendor Id: 0x1030 > Short Name: TMC > > Does "fmdump -eV" give any clue when the box comes back up? > > ..Remco > > > Blake wrote: >> >> I''m attaching a screenshot of the console just before reboot. ?The >> dump doesn''t seem to be working, or savecore isn''t working. >> >> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>> >>> I''m working on testing this some more by doing a savecore -L right >>> after I start the copy. >>> >>> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >>> supports 16, but isn''t fully populated). >>> >>> So far as I know, there is no spinup happening - these are not RAID >>> controllers, just dumb SATA JBOD controllers, so I don''t think they >>> control drive spin in any particular way. ?Correct me if I''m wrong, of >>> course. >>> >>> >>> >>> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >>>> >>>> The copy operation will make all the disks start seeking at the same >>>> time and >>>> will make your CPU activity jump to a significant percentage to compute >>>> the >>>> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >>>> because of the sudden increase in power consumption... >>>> >>>> However if you are *not* using SATA staggered spin-up, then the above >>>> theory >>>> is unlikely because spinning up consumes much more power than when >>>> seeking. >>>> So, in a sense, a successful boot proves your PSU is powerful enough. >>>> >>>> Trying reproducing the problem by copying data on a smaller number of >>>> disks. >>>> You tried 2 and 16. Try 8. >>>> >>>> -marc >>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>> >>> >>> ------------------------------------------------------------------------ >>> >>> >>> ------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >
Could the problem be related to this bug: http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6793353 I''m testing setting the maximum payload size as a workaround, as noted in the bug notes. On Wed, Mar 11, 2009 at 3:14 PM, Blake <blake.irvin at gmail.com> wrote:> I think that TMC Research is the company that designed the > Supermicro-branded controller card that has the Marvell SATA > controller chip on it. ?Googling around I see connections between > Supermicro and TMC. > > This is the card: > > http://www.supermicro.com/products/accessories/addon/AOC-SAT2-MV8.cfm > > On Wed, Mar 11, 2009 at 3:08 PM, Remco Lengers <remco at lengers.com> wrote: >> Something is not right in the IO space. The messages talk about >> >> vendor ID = 11AB >> >> 0x11AB ?Marvell Semiconductor >> >> TMC Research >> >> Vendor Id: 0x1030 >> Short Name: TMC >> >> Does "fmdump -eV" give any clue when the box comes back up? >> >> ..Remco >> >> >> Blake wrote: >>> >>> I''m attaching a screenshot of the console just before reboot. ?The >>> dump doesn''t seem to be working, or savecore isn''t working. >>> >>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>> >>>> I''m working on testing this some more by doing a savecore -L right >>>> after I start the copy. >>>> >>>> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >>>> supports 16, but isn''t fully populated). >>>> >>>> So far as I know, there is no spinup happening - these are not RAID >>>> controllers, just dumb SATA JBOD controllers, so I don''t think they >>>> control drive spin in any particular way. ?Correct me if I''m wrong, of >>>> course. >>>> >>>> >>>> >>>> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >>>>> >>>>> The copy operation will make all the disks start seeking at the same >>>>> time and >>>>> will make your CPU activity jump to a significant percentage to compute >>>>> the >>>>> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >>>>> because of the sudden increase in power consumption... >>>>> >>>>> However if you are *not* using SATA staggered spin-up, then the above >>>>> theory >>>>> is unlikely because spinning up consumes much more power than when >>>>> seeking. >>>>> So, in a sense, a successful boot proves your PSU is powerful enough. >>>>> >>>>> Trying reproducing the problem by copying data on a smaller number of >>>>> disks. >>>>> You tried 2 and 16. Try 8. >>>>> >>>>> -marc >>>>> >>>>> _______________________________________________ >>>>> zfs-discuss mailing list >>>>> zfs-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>>> >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> >
Remco Lengers
2009-Mar-11 19:41 UTC
[zfs-discuss] reboot when copying large amounts of data
looks worth a go otherwise: if the boot disk is also off that controller it may be too hosed to write anything to the boot disk hence FMA doesn''t see any issue when it comes up. Possible further actions: - Upgrade FW of controller to highest or known working level - Upgrade driver or OS level. - Try another controller (may be its broken and barfs under stress ?) - Analyze the crash dump (if any is saved) - It may be its a know Solaris or driver bug and somebody has heard of it before. hth, ..Remco Blake wrote:> Could the problem be related to this bug: > > http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6793353 > > I''m testing setting the maximum payload size as a workaround, as noted > in the bug notes. > > > > On Wed, Mar 11, 2009 at 3:14 PM, Blake <blake.irvin at gmail.com> wrote: >> I think that TMC Research is the company that designed the >> Supermicro-branded controller card that has the Marvell SATA >> controller chip on it. Googling around I see connections between >> Supermicro and TMC. >> >> This is the card: >> >> http://www.supermicro.com/products/accessories/addon/AOC-SAT2-MV8.cfm >> >> On Wed, Mar 11, 2009 at 3:08 PM, Remco Lengers <remco at lengers.com> wrote: >>> Something is not right in the IO space. The messages talk about >>> >>> vendor ID = 11AB >>> >>> 0x11AB Marvell Semiconductor >>> >>> TMC Research >>> >>> Vendor Id: 0x1030 >>> Short Name: TMC >>> >>> Does "fmdump -eV" give any clue when the box comes back up? >>> >>> ..Remco >>> >>> >>> Blake wrote: >>>> I''m attaching a screenshot of the console just before reboot. The >>>> dump doesn''t seem to be working, or savecore isn''t working. >>>> >>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>>> I''m working on testing this some more by doing a savecore -L right >>>>> after I start the copy. >>>>> >>>>> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >>>>> supports 16, but isn''t fully populated). >>>>> >>>>> So far as I know, there is no spinup happening - these are not RAID >>>>> controllers, just dumb SATA JBOD controllers, so I don''t think they >>>>> control drive spin in any particular way. Correct me if I''m wrong, of >>>>> course. >>>>> >>>>> >>>>> >>>>> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> wrote: >>>>>> The copy operation will make all the disks start seeking at the same >>>>>> time and >>>>>> will make your CPU activity jump to a significant percentage to compute >>>>>> the >>>>>> ZFS checksum and RAIDZ parity. I think you could be overloading your PSU >>>>>> because of the sudden increase in power consumption... >>>>>> >>>>>> However if you are *not* using SATA staggered spin-up, then the above >>>>>> theory >>>>>> is unlikely because spinning up consumes much more power than when >>>>>> seeking. >>>>>> So, in a sense, a successful boot proves your PSU is powerful enough. >>>>>> >>>>>> Trying reproducing the problem by copying data on a smaller number of >>>>>> disks. >>>>>> You tried 2 and 16. Try 8. >>>>>> >>>>>> -marc >>>>>> >>>>>> _______________________________________________ >>>>>> zfs-discuss mailing list >>>>>> zfs-discuss at opensolaris.org >>>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>>>> >>>>> ------------------------------------------------------------------------ >>>>> >>>>> >>>>> ------------------------------------------------------------------------ >>>>> >>>>> _______________________________________________ >>>>> zfs-discuss mailing list >>>>> zfs-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Any chance this could be the motherboard? I suspect the controller. The boot disks are on the built-in nVidia controller. On Wed, Mar 11, 2009 at 3:41 PM, Remco Lengers <remco at lengers.com> wrote:> - Upgrade FW of controller to highest or known working levelI think I have the latest controller firmware.> - Upgrade driver or OS level.I''m going to try to go from 101b to 108 or whatever the current dev release is.> - Try another controller (may be its broken and barfs under stress ?)In the works.> - Analyze the crash dump (if any is saved)Crash dump is not saving properly.> - It may be its a know Solaris or driver bug and somebody has heard of it > before.Any takers on this? :)> > hth,Thanks!> > ..Remco > > Blake wrote: >> >> Could the problem be related to this bug: >> >> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6793353 >> >> I''m testing setting the maximum payload size as a workaround, as noted >> in the bug notes. >> >> >> >> On Wed, Mar 11, 2009 at 3:14 PM, Blake <blake.irvin at gmail.com> wrote: >>> >>> I think that TMC Research is the company that designed the >>> Supermicro-branded controller card that has the Marvell SATA >>> controller chip on it. ?Googling around I see connections between >>> Supermicro and TMC. >>> >>> This is the card: >>> >>> http://www.supermicro.com/products/accessories/addon/AOC-SAT2-MV8.cfm >>> >>> On Wed, Mar 11, 2009 at 3:08 PM, Remco Lengers <remco at lengers.com> wrote: >>>> >>>> Something is not right in the IO space. The messages talk about >>>> >>>> vendor ID = 11AB >>>> >>>> 0x11AB ?Marvell Semiconductor >>>> >>>> TMC Research >>>> >>>> Vendor Id: 0x1030 >>>> Short Name: TMC >>>> >>>> Does "fmdump -eV" give any clue when the box comes back up? >>>> >>>> ..Remco >>>> >>>> >>>> Blake wrote: >>>>> >>>>> I''m attaching a screenshot of the console just before reboot. ?The >>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>> >>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>>>> >>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>> after I start the copy. >>>>>> >>>>>> BTW, I''m copying to a raidz2 of only 5 disks, not 16 (the chassis >>>>>> supports 16, but isn''t fully populated). >>>>>> >>>>>> So far as I know, there is no spinup happening - these are not RAID >>>>>> controllers, just dumb SATA JBOD controllers, so I don''t think they >>>>>> control drive spin in any particular way. ?Correct me if I''m wrong, of >>>>>> course. >>>>>> >>>>>> >>>>>> >>>>>> On Wed, Mar 11, 2009 at 11:23 AM, Marc Bevand <m.bevand at gmail.com> >>>>>> wrote: >>>>>>> >>>>>>> The copy operation will make all the disks start seeking at the same >>>>>>> time and >>>>>>> will make your CPU activity jump to a significant percentage to >>>>>>> compute >>>>>>> the >>>>>>> ZFS checksum and RAIDZ parity. I think you could be overloading your >>>>>>> PSU >>>>>>> because of the sudden increase in power consumption... >>>>>>> >>>>>>> However if you are *not* using SATA staggered spin-up, then the above >>>>>>> theory >>>>>>> is unlikely because spinning up consumes much more power than when >>>>>>> seeking. >>>>>>> So, in a sense, a successful boot proves your PSU is powerful enough. >>>>>>> >>>>>>> Trying reproducing the problem by copying data on a smaller number of >>>>>>> disks. >>>>>>> You tried 2 and 16. Try 8. >>>>>>> >>>>>>> -marc >>>>>>> >>>>>>> _______________________________________________ >>>>>>> zfs-discuss mailing list >>>>>>> zfs-discuss at opensolaris.org >>>>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------ >>>>>> >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------ >>>>>> >>>>>> _______________________________________________ >>>>>> zfs-discuss mailing list >>>>>> zfs-discuss at opensolaris.org >>>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >
Richard Elling
2009-Mar-11 20:58 UTC
[zfs-discuss] reboot when copying large amounts of data
Blake wrote:> I''m attaching a screenshot of the console just before reboot. The > dump doesn''t seem to be working, or savecore isn''t working. > > On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: > >> I''m working on testing this some more by doing a savecore -L right >> after I start the copy. >> >>savecore -L is not what you want. By default, for OpenSolaris, savecore on boot is disabled. But the core will have been dumped into the dump slice, which is not used for swap. So you should be able to run savecore at a later time to collect the core from the last dump. -- richard
I guess I didn''t make it clear that I had already tried using savecore to retrieve the core from the dump device. I added a larger zvol for dump, to make sure that I wasn''t running out of space on the dump device: root at host:~# dumpadm Dump content: kernel pages Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore directory: /var/crash/host Savecore enabled: yes I was using the -L option only to try to get some idea of why the system load was climbing to 1 during a simple file copy. On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling <richard.elling at gmail.com> wrote:> Blake wrote: >> >> I''m attaching a screenshot of the console just before reboot. ?The >> dump doesn''t seem to be working, or savecore isn''t working. >> >> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >> >>> >>> I''m working on testing this some more by doing a savecore -L right >>> after I start the copy. >>> >>> > > savecore -L is not what you want. > > By default, for OpenSolaris, savecore on boot is disabled. ?But the core > will have been dumped into the dump slice, which is not used for swap. > So you should be able to run savecore at a later time to collect the > core from the last dump. > -- richard > >
Maidak Alexander J
2009-Mar-12 04:31 UTC
[zfs-discuss] reboot when copying large amounts of data
If you''re having issues with a disk contoller or disk IO driver its highly likely that a savecore to disk after the panic will fail. I''m not sure how to work around this, maybe a dedicated dump device not on a controller that uses a different driver then the one that you''re having issues with? -----Original Message----- From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake Sent: Wednesday, March 11, 2009 4:45 PM To: Richard Elling Cc: Marc Bevand; zfs-discuss at opensolaris.org Subject: Re: [zfs-discuss] reboot when copying large amounts of data I guess I didn''t make it clear that I had already tried using savecore to retrieve the core from the dump device. I added a larger zvol for dump, to make sure that I wasn''t running out of space on the dump device: root at host:~# dumpadm Dump content: kernel pages Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore directory: /var/crash/host Savecore enabled: yes I was using the -L option only to try to get some idea of why the system load was climbing to 1 during a simple file copy. On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling <richard.elling at gmail.com> wrote:> Blake wrote: >> >> I''m attaching a screenshot of the console just before reboot. ?The >> dump doesn''t seem to be working, or savecore isn''t working. >> >> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >> >>> >>> I''m working on testing this some more by doing a savecore -L right >>> after I start the copy. >>> >>> > > savecore -L is not what you want. > > By default, for OpenSolaris, savecore on boot is disabled. ?But the > core will have been dumped into the dump slice, which is not used for swap. > So you should be able to run savecore at a later time to collect the > core from the last dump. > -- richard > >_______________________________________________ zfs-discuss mailing list zfs-discuss at opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
My dump device is already on a different controller - the motherboards built-in nVidia SATA controller. The raidz2 vdev is the one I''m having trouble with (copying the same files to the mirrored rpool on the nVidia controller work nicely). I do notice that, when using cp to copy the files to the raidz2 pool, load on the machine climbs steadily until the crash, and one proc core pegs at 100%. Frustrating, yes. On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J <MaidakAlexanderJ at johndeere.com> wrote:> If you''re having issues with a disk contoller or disk IO driver its highly likely that a savecore to disk after the panic will fail. ?I''m not sure how to work around this, maybe a dedicated dump device not on a controller that uses a different driver then the one that you''re having issues with? > > -----Original Message----- > From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake > Sent: Wednesday, March 11, 2009 4:45 PM > To: Richard Elling > Cc: Marc Bevand; zfs-discuss at opensolaris.org > Subject: Re: [zfs-discuss] reboot when copying large amounts of data > > I guess I didn''t make it clear that I had already tried using savecore to retrieve the core from the dump device. > > I added a larger zvol for dump, to make sure that I wasn''t running out of space on the dump device: > > root at host:~# dumpadm > ? ? ?Dump content: kernel pages > ? ? ? Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore directory: /var/crash/host > ?Savecore enabled: yes > > I was using the -L option only to try to get some idea of why the system load was climbing to 1 during a simple file copy. > > > > On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling <richard.elling at gmail.com> wrote: >> Blake wrote: >>> >>> I''m attaching a screenshot of the console just before reboot. ?The >>> dump doesn''t seem to be working, or savecore isn''t working. >>> >>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>> >>>> >>>> I''m working on testing this some more by doing a savecore -L right >>>> after I start the copy. >>>> >>>> >> >> savecore -L is not what you want. >> >> By default, for OpenSolaris, savecore on boot is disabled. ?But the >> core will have been dumped into the dump slice, which is not used for swap. >> So you should be able to run savecore at a later time to collect the >> core from the last dump. >> -- richard >> >> > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >
Nathan Kroenert
2009-Mar-12 05:55 UTC
[zfs-discuss] reboot when copying large amounts of data
Hm - Crashes, or hangs? Moreover - how do you know a CPU is pegged? Seems like we could do a little more discovery on what the actual problem here is, as I can read it about 4 different ways. By this last piece of information, I''m guessing the system does not crash, but goes really really slow?? Crash == panic == we see stack dump on console and try to take a dump hang == nothing works == no response -> might be worth looking at mdb -K or booting with a -k on the boot line. So - are we crashing, hanging, or something different? It might simply be that you are eating up all your memory, and your physical backing storage is taking a while to catch up....? Nathan. Blake wrote:> My dump device is already on a different controller - the motherboards > built-in nVidia SATA controller. > > The raidz2 vdev is the one I''m having trouble with (copying the same > files to the mirrored rpool on the nVidia controller work nicely). I > do notice that, when using cp to copy the files to the raidz2 pool, > load on the machine climbs steadily until the crash, and one proc core > pegs at 100%. > > Frustrating, yes. > > On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J > <MaidakAlexanderJ at johndeere.com> wrote: >> If you''re having issues with a disk contoller or disk IO driver its highly likely that a savecore to disk after the panic will fail. I''m not sure how to work around this, maybe a dedicated dump device not on a controller that uses a different driver then the one that you''re having issues with? >> >> -----Original Message----- >> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >> Sent: Wednesday, March 11, 2009 4:45 PM >> To: Richard Elling >> Cc: Marc Bevand; zfs-discuss at opensolaris.org >> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >> >> I guess I didn''t make it clear that I had already tried using savecore to retrieve the core from the dump device. >> >> I added a larger zvol for dump, to make sure that I wasn''t running out of space on the dump device: >> >> root at host:~# dumpadm >> Dump content: kernel pages >> Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore directory: /var/crash/host >> Savecore enabled: yes >> >> I was using the -L option only to try to get some idea of why the system load was climbing to 1 during a simple file copy. >> >> >> >> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling <richard.elling at gmail.com> wrote: >>> Blake wrote: >>>> I''m attaching a screenshot of the console just before reboot. The >>>> dump doesn''t seem to be working, or savecore isn''t working. >>>> >>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>> >>>>> I''m working on testing this some more by doing a savecore -L right >>>>> after I start the copy. >>>>> >>>>> >>> savecore -L is not what you want. >>> >>> By default, for OpenSolaris, savecore on boot is disabled. But the >>> core will have been dumped into the dump slice, which is not used for swap. >>> So you should be able to run savecore at a later time to collect the >>> core from the last dump. >>> -- richard >>> >>> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss-- ////////////////////////////////////////////////////////////////// // Nathan Kroenert nathan.kroenert at sun.com // // Systems Engineer Phone: +61 3 9869-6255 // // Sun Microsystems Fax: +61 3 9869-6288 // // Level 7, 476 St. Kilda Road Mobile: 0419 305 456 // // Melbourne 3004 Victoria Australia // //////////////////////////////////////////////////////////////////
I start the cp, and then, with prstat -a, watch the cpu load for the cp process climb to 25% on a 4-core machine. Load, measured for example with ''uptime'', climbs steadily until the reboot. Note that the machine does not dump properly, panic or hang - rather, it reboots. I attached a screenshot earlier in this thread of the little bit of error message I could see on the console. The machine is trying to dump to the dump zvol, but fails to do so. Only sometimes do I see an error on the machine''s local console - mos times, it simply reboots. On Thu, Mar 12, 2009 at 1:55 AM, Nathan Kroenert <Nathan.Kroenert at sun.com> wrote:> Hm - > > Crashes, or hangs? Moreover - how do you know a CPU is pegged? > > Seems like we could do a little more discovery on what the actual problem > here is, as I can read it about 4 different ways. > > By this last piece of information, I''m guessing the system does not crash, > but goes really really slow?? > > Crash == panic == we see stack dump on console and try to take a dump > hang == nothing works == no response -> might be worth looking at mdb -K > ? ? ? ?or booting with a -k on the boot line. > > So - are we crashing, hanging, or something different? > > It might simply be that you are eating up all your memory, and your physical > backing storage is taking a while to catch up....? > > Nathan. > > Blake wrote: >> >> My dump device is already on a different controller - the motherboards >> built-in nVidia SATA controller. >> >> The raidz2 vdev is the one I''m having trouble with (copying the same >> files to the mirrored rpool on the nVidia controller work nicely). ?I >> do notice that, when using cp to copy the files to the raidz2 pool, >> load on the machine climbs steadily until the crash, and one proc core >> pegs at 100%. >> >> Frustrating, yes. >> >> On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J >> <MaidakAlexanderJ at johndeere.com> wrote: >>> >>> If you''re having issues with a disk contoller or disk IO driver its >>> highly likely that a savecore to disk after the panic will fail. ?I''m not >>> sure how to work around this, maybe a dedicated dump device not on a >>> controller that uses a different driver then the one that you''re having >>> issues with? >>> >>> -----Original Message----- >>> From: zfs-discuss-bounces at opensolaris.org >>> [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >>> Sent: Wednesday, March 11, 2009 4:45 PM >>> To: Richard Elling >>> Cc: Marc Bevand; zfs-discuss at opensolaris.org >>> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >>> >>> I guess I didn''t make it clear that I had already tried using savecore to >>> retrieve the core from the dump device. >>> >>> I added a larger zvol for dump, to make sure that I wasn''t running out of >>> space on the dump device: >>> >>> root at host:~# dumpadm >>> ? ? Dump content: kernel pages >>> ? ? ?Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore >>> directory: /var/crash/host >>> ?Savecore enabled: yes >>> >>> I was using the -L option only to try to get some idea of why the system >>> load was climbing to 1 during a simple file copy. >>> >>> >>> >>> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling >>> <richard.elling at gmail.com> wrote: >>>> >>>> Blake wrote: >>>>> >>>>> I''m attaching a screenshot of the console just before reboot. ?The >>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>> >>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>>> >>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>> after I start the copy. >>>>>> >>>>>> >>>> savecore -L is not what you want. >>>> >>>> By default, for OpenSolaris, savecore on boot is disabled. ?But the >>>> core will have been dumped into the dump slice, which is not used for >>>> swap. >>>> So you should be able to run savecore at a later time to collect the >>>> core from the last dump. >>>> -- richard >>>> >>>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss > > -- > ////////////////////////////////////////////////////////////////// > // Nathan Kroenert ? ? ? ? ? ? ?nathan.kroenert at sun.com ? ? ? ? // > // Systems Engineer ? ? ? ? ? ? Phone: ?+61 3 9869-6255 ? ? ? ? // > // Sun Microsystems ? ? ? ? ? ? Fax: ? ?+61 3 9869-6288 ? ? ? ? // > // Level 7, 476 St. Kilda Road ?Mobile: 0419 305 456 ? ? ? ? ? ?// > // Melbourne 3004 ? Victoria ? ?Australia ? ? ? ? ? ? ? ? ? ? ? // > ////////////////////////////////////////////////////////////////// >
Nathan Kroenert
2009-Mar-12 06:18 UTC
[zfs-discuss] reboot when copying large amounts of data
definitely time to bust out some mdb -k and see what it''s moaning about. I did not see the screenshot earlier... sorry about that. Nathan. Blake wrote:> I start the cp, and then, with prstat -a, watch the cpu load for the > cp process climb to 25% on a 4-core machine. > > Load, measured for example with ''uptime'', climbs steadily until the reboot. > > Note that the machine does not dump properly, panic or hang - rather, > it reboots. > > I attached a screenshot earlier in this thread of the little bit of > error message I could see on the console. The machine is trying to > dump to the dump zvol, but fails to do so. Only sometimes do I see an > error on the machine''s local console - mos times, it simply reboots. > > > > On Thu, Mar 12, 2009 at 1:55 AM, Nathan Kroenert > <Nathan.Kroenert at sun.com> wrote: >> Hm - >> >> Crashes, or hangs? Moreover - how do you know a CPU is pegged? >> >> Seems like we could do a little more discovery on what the actual problem >> here is, as I can read it about 4 different ways. >> >> By this last piece of information, I''m guessing the system does not crash, >> but goes really really slow?? >> >> Crash == panic == we see stack dump on console and try to take a dump >> hang == nothing works == no response -> might be worth looking at mdb -K >> or booting with a -k on the boot line. >> >> So - are we crashing, hanging, or something different? >> >> It might simply be that you are eating up all your memory, and your physical >> backing storage is taking a while to catch up....? >> >> Nathan. >> >> Blake wrote: >>> My dump device is already on a different controller - the motherboards >>> built-in nVidia SATA controller. >>> >>> The raidz2 vdev is the one I''m having trouble with (copying the same >>> files to the mirrored rpool on the nVidia controller work nicely). I >>> do notice that, when using cp to copy the files to the raidz2 pool, >>> load on the machine climbs steadily until the crash, and one proc core >>> pegs at 100%. >>> >>> Frustrating, yes. >>> >>> On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J >>> <MaidakAlexanderJ at johndeere.com> wrote: >>>> If you''re having issues with a disk contoller or disk IO driver its >>>> highly likely that a savecore to disk after the panic will fail. I''m not >>>> sure how to work around this, maybe a dedicated dump device not on a >>>> controller that uses a different driver then the one that you''re having >>>> issues with? >>>> >>>> -----Original Message----- >>>> From: zfs-discuss-bounces at opensolaris.org >>>> [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >>>> Sent: Wednesday, March 11, 2009 4:45 PM >>>> To: Richard Elling >>>> Cc: Marc Bevand; zfs-discuss at opensolaris.org >>>> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >>>> >>>> I guess I didn''t make it clear that I had already tried using savecore to >>>> retrieve the core from the dump device. >>>> >>>> I added a larger zvol for dump, to make sure that I wasn''t running out of >>>> space on the dump device: >>>> >>>> root at host:~# dumpadm >>>> Dump content: kernel pages >>>> Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore >>>> directory: /var/crash/host >>>> Savecore enabled: yes >>>> >>>> I was using the -L option only to try to get some idea of why the system >>>> load was climbing to 1 during a simple file copy. >>>> >>>> >>>> >>>> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling >>>> <richard.elling at gmail.com> wrote: >>>>> Blake wrote: >>>>>> I''m attaching a screenshot of the console just before reboot. The >>>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>>> >>>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>>>> >>>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>>> after I start the copy. >>>>>>> >>>>>>> >>>>> savecore -L is not what you want. >>>>> >>>>> By default, for OpenSolaris, savecore on boot is disabled. But the >>>>> core will have been dumped into the dump slice, which is not used for >>>>> swap. >>>>> So you should be able to run savecore at a later time to collect the >>>>> core from the last dump. >>>>> -- richard >>>>> >>>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> -- >> ////////////////////////////////////////////////////////////////// >> // Nathan Kroenert nathan.kroenert at sun.com // >> // Systems Engineer Phone: +61 3 9869-6255 // >> // Sun Microsystems Fax: +61 3 9869-6288 // >> // Level 7, 476 St. Kilda Road Mobile: 0419 305 456 // >> // Melbourne 3004 Victoria Australia // >> ////////////////////////////////////////////////////////////////// >>-- ////////////////////////////////////////////////////////////////// // Nathan Kroenert nathan.kroenert at sun.com // // Systems Engineer Phone: +61 3 9869-6255 // // Sun Microsystems Fax: +61 3 9869-6288 // // Level 7, 476 St. Kilda Road Mobile: 0419 305 456 // // Melbourne 3004 Victoria Australia // //////////////////////////////////////////////////////////////////
Nathan Kroenert
2009-Mar-12 06:18 UTC
[zfs-discuss] reboot when copying large amounts of data
definitely time to bust out some mdb -K or boot -k and see what it''s moaning about. I did not see the screenshot earlier... sorry about that. Nathan. Blake wrote:> I start the cp, and then, with prstat -a, watch the cpu load for the > cp process climb to 25% on a 4-core machine. > > Load, measured for example with ''uptime'', climbs steadily until the reboot. > > Note that the machine does not dump properly, panic or hang - rather, > it reboots. > > I attached a screenshot earlier in this thread of the little bit of > error message I could see on the console. The machine is trying to > dump to the dump zvol, but fails to do so. Only sometimes do I see an > error on the machine''s local console - mos times, it simply reboots. > > > > On Thu, Mar 12, 2009 at 1:55 AM, Nathan Kroenert > <Nathan.Kroenert at sun.com> wrote: >> Hm - >> >> Crashes, or hangs? Moreover - how do you know a CPU is pegged? >> >> Seems like we could do a little more discovery on what the actual problem >> here is, as I can read it about 4 different ways. >> >> By this last piece of information, I''m guessing the system does not crash, >> but goes really really slow?? >> >> Crash == panic == we see stack dump on console and try to take a dump >> hang == nothing works == no response -> might be worth looking at mdb -K >> or booting with a -k on the boot line. >> >> So - are we crashing, hanging, or something different? >> >> It might simply be that you are eating up all your memory, and your physical >> backing storage is taking a while to catch up....? >> >> Nathan. >> >> Blake wrote: >>> My dump device is already on a different controller - the motherboards >>> built-in nVidia SATA controller. >>> >>> The raidz2 vdev is the one I''m having trouble with (copying the same >>> files to the mirrored rpool on the nVidia controller work nicely). I >>> do notice that, when using cp to copy the files to the raidz2 pool, >>> load on the machine climbs steadily until the crash, and one proc core >>> pegs at 100%. >>> >>> Frustrating, yes. >>> >>> On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J >>> <MaidakAlexanderJ at johndeere.com> wrote: >>>> If you''re having issues with a disk contoller or disk IO driver its >>>> highly likely that a savecore to disk after the panic will fail. I''m not >>>> sure how to work around this, maybe a dedicated dump device not on a >>>> controller that uses a different driver then the one that you''re having >>>> issues with? >>>> >>>> -----Original Message----- >>>> From: zfs-discuss-bounces at opensolaris.org >>>> [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >>>> Sent: Wednesday, March 11, 2009 4:45 PM >>>> To: Richard Elling >>>> Cc: Marc Bevand; zfs-discuss at opensolaris.org >>>> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >>>> >>>> I guess I didn''t make it clear that I had already tried using savecore to >>>> retrieve the core from the dump device. >>>> >>>> I added a larger zvol for dump, to make sure that I wasn''t running out of >>>> space on the dump device: >>>> >>>> root at host:~# dumpadm >>>> Dump content: kernel pages >>>> Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore >>>> directory: /var/crash/host >>>> Savecore enabled: yes >>>> >>>> I was using the -L option only to try to get some idea of why the system >>>> load was climbing to 1 during a simple file copy. >>>> >>>> >>>> >>>> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling >>>> <richard.elling at gmail.com> wrote: >>>>> Blake wrote: >>>>>> I''m attaching a screenshot of the console just before reboot. The >>>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>>> >>>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> wrote: >>>>>> >>>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>>> after I start the copy. >>>>>>> >>>>>>> >>>>> savecore -L is not what you want. >>>>> >>>>> By default, for OpenSolaris, savecore on boot is disabled. But the >>>>> core will have been dumped into the dump slice, which is not used for >>>>> swap. >>>>> So you should be able to run savecore at a later time to collect the >>>>> core from the last dump. >>>>> -- richard >>>>> >>>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>> >>> _______________________________________________ >>> zfs-discuss mailing list >>> zfs-discuss at opensolaris.org >>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> -- >> ////////////////////////////////////////////////////////////////// >> // Nathan Kroenert nathan.kroenert at sun.com // >> // Systems Engineer Phone: +61 3 9869-6255 // >> // Sun Microsystems Fax: +61 3 9869-6288 // >> // Level 7, 476 St. Kilda Road Mobile: 0419 305 456 // >> // Melbourne 3004 Victoria Australia // >> ////////////////////////////////////////////////////////////////// >>-- ////////////////////////////////////////////////////////////////// // Nathan Kroenert nathan.kroenert at sun.com // // Systems Engineer Phone: +61 3 9869-6255 // // Sun Microsystems Fax: +61 3 9869-6288 // // Level 7, 476 St. Kilda Road Mobile: 0419 305 456 // // Melbourne 3004 Victoria Australia // //////////////////////////////////////////////////////////////////
So, if I boot with the -k boot flags (to load the kernel debugger?) what do I need to look for? I''m no expert at kernel debugging. I think this is a pci error judging by the console output, or at least is i/o related... thanks for your feedback, Blake On Thu, Mar 12, 2009 at 2:18 AM, Nathan Kroenert <Nathan.Kroenert at sun.com> wrote:> definitely time to bust out some mdb -K or boot -k and see what it''s moaning > about. > > I did not see the screenshot earlier... sorry about that. > > Nathan. > > Blake wrote: >> >> I start the cp, and then, with prstat -a, watch the cpu load for the >> cp process climb to 25% on a 4-core machine. >> >> Load, measured for example with ''uptime'', climbs steadily until the >> reboot. >> >> Note that the machine does not dump properly, panic or hang - rather, >> it reboots. >> >> I attached a screenshot earlier in this thread of the little bit of >> error message I could see on the console. ?The machine is trying to >> dump to the dump zvol, but fails to do so. ?Only sometimes do I see an >> error on the machine''s local console - mos times, it simply reboots. >> >> >> >> On Thu, Mar 12, 2009 at 1:55 AM, Nathan Kroenert >> <Nathan.Kroenert at sun.com> wrote: >>> >>> Hm - >>> >>> Crashes, or hangs? Moreover - how do you know a CPU is pegged? >>> >>> Seems like we could do a little more discovery on what the actual problem >>> here is, as I can read it about 4 different ways. >>> >>> By this last piece of information, I''m guessing the system does not >>> crash, >>> but goes really really slow?? >>> >>> Crash == panic == we see stack dump on console and try to take a dump >>> hang == nothing works == no response -> might be worth looking at mdb -K >>> ? ? ? or booting with a -k on the boot line. >>> >>> So - are we crashing, hanging, or something different? >>> >>> It might simply be that you are eating up all your memory, and your >>> physical >>> backing storage is taking a while to catch up....? >>> >>> Nathan. >>> >>> Blake wrote: >>>> >>>> My dump device is already on a different controller - the motherboards >>>> built-in nVidia SATA controller. >>>> >>>> The raidz2 vdev is the one I''m having trouble with (copying the same >>>> files to the mirrored rpool on the nVidia controller work nicely). ?I >>>> do notice that, when using cp to copy the files to the raidz2 pool, >>>> load on the machine climbs steadily until the crash, and one proc core >>>> pegs at 100%. >>>> >>>> Frustrating, yes. >>>> >>>> On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J >>>> <MaidakAlexanderJ at johndeere.com> wrote: >>>>> >>>>> If you''re having issues with a disk contoller or disk IO driver its >>>>> highly likely that a savecore to disk after the panic will fail. ?I''m >>>>> not >>>>> sure how to work around this, maybe a dedicated dump device not on a >>>>> controller that uses a different driver then the one that you''re having >>>>> issues with? >>>>> >>>>> -----Original Message----- >>>>> From: zfs-discuss-bounces at opensolaris.org >>>>> [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >>>>> Sent: Wednesday, March 11, 2009 4:45 PM >>>>> To: Richard Elling >>>>> Cc: Marc Bevand; zfs-discuss at opensolaris.org >>>>> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >>>>> >>>>> I guess I didn''t make it clear that I had already tried using savecore >>>>> to >>>>> retrieve the core from the dump device. >>>>> >>>>> I added a larger zvol for dump, to make sure that I wasn''t running out >>>>> of >>>>> space on the dump device: >>>>> >>>>> root at host:~# dumpadm >>>>> ? ?Dump content: kernel pages >>>>> ? ? Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore >>>>> directory: /var/crash/host >>>>> ?Savecore enabled: yes >>>>> >>>>> I was using the -L option only to try to get some idea of why the >>>>> system >>>>> load was climbing to 1 during a simple file copy. >>>>> >>>>> >>>>> >>>>> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling >>>>> <richard.elling at gmail.com> wrote: >>>>>> >>>>>> Blake wrote: >>>>>>> >>>>>>> I''m attaching a screenshot of the console just before reboot. ?The >>>>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>>>> >>>>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>>>> after I start the copy. >>>>>>>> >>>>>>>> >>>>>> savecore -L is not what you want. >>>>>> >>>>>> By default, for OpenSolaris, savecore on boot is disabled. ?But the >>>>>> core will have been dumped into the dump slice, which is not used for >>>>>> swap. >>>>>> So you should be able to run savecore at a later time to collect the >>>>>> core from the last dump. >>>>>> -- richard >>>>>> >>>>>> >>>>> _______________________________________________ >>>>> zfs-discuss mailing list >>>>> zfs-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>> >>> -- >>> ////////////////////////////////////////////////////////////////// >>> // Nathan Kroenert ? ? ? ? ? ? ?nathan.kroenert at sun.com ? ? ? ? // >>> // Systems Engineer ? ? ? ? ? ? Phone: ?+61 3 9869-6255 ? ? ? ? // >>> // Sun Microsystems ? ? ? ? ? ? Fax: ? ?+61 3 9869-6288 ? ? ? ? // >>> // Level 7, 476 St. Kilda Road ?Mobile: 0419 305 456 ? ? ? ? ? ?// >>> // Melbourne 3004 ? Victoria ? ?Australia ? ? ? ? ? ? ? ? ? ? ? // >>> ////////////////////////////////////////////////////////////////// >>> > > -- > ////////////////////////////////////////////////////////////////// > // Nathan Kroenert ? ? ? ? ? ? ?nathan.kroenert at sun.com ? ? ? ? // > // Systems Engineer ? ? ? ? ? ? Phone: ?+61 3 9869-6255 ? ? ? ? // > // Sun Microsystems ? ? ? ? ? ? Fax: ? ?+61 3 9869-6288 ? ? ? ? // > // Level 7, 476 St. Kilda Road ?Mobile: 0419 305 456 ? ? ? ? ? ?// > // Melbourne 3004 ? Victoria ? ?Australia ? ? ? ? ? ? ? ? ? ? ? // > ////////////////////////////////////////////////////////////////// >
Miles Nordin
2009-Mar-12 18:12 UTC
[zfs-discuss] reboot when copying large amounts of data
>>>>> "maj" == Maidak Alexander J <MaidakAlexanderJ at JohnDeere.com> writes:maj> If you''re having issues with a disk contoller or disk IO maj> driver its highly likely that a savecore to disk after the maj> panic will fail. I''m not sure how to work around this not in Solaris, but as a concept for solving the problem: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=Documentation/kdump/kdump.txt;h=3f4bc840da8b7c068076dd057216e846e098db9f;hb=4a6908a3a050aacc9c3a2f36b276b46c0629ad91 They load a second kernel into a reserved spot of RAM, like 64MB or so, and forget about it. After a crash, they boot the second kernel. The second kernel runs using the reserved area of RAM as its working space, not touching any other memory, as if you were running on a very old machine with tiny RAM. It reprobes all the hardware, and then performs the dump. I don''t know if it actually works, but the approach is appropriate if you are trying to debug the storage stack. You could even have a main kernel which crashes while taking an ordinary coredump, and then use the backup dumping-kernel to coredump the main kernel in mid-coredump---a dump of a dumping kernel. I think some Solaris developers were discussing putting coredump features into Xen, so the host could take the dump (or, maybe even something better than a dump---for example, if you built host/target debugging features into Xen for debugging running kernels, then you could just force a breakpoint in the guest instead of panic. Since Xen can hibernate domU''s onto disk (it can, right?), you can treat the hibernated Xen-specific representation of the domU as the-dump, groveling through the ``dump'''' with the same host/target tools you could use on a running kernel without any special dump support in the debugger itself). IIRC NetBSD developers discussed the same idea years ago but neither implementation exists. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 304 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090312/0d730cae/attachment.bin>
I''ve managed to get the data transfer to work by rearranging my disks so that all of them sit on the integrated SATA controller. So, I feel pretty certain that this is either an issue with the Supermicro aoc-sat2-mv8 card, or with PCI-X on the motherboard (though I would think that the integrated SATA would also be using the PCI bus?). The motherboard, for those interested, is an HD8ME-2 (not, I now find after buying this box from Silicon Mechanics, a board that''s on the Solaris HCL...) <http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/h8dme-2.cfm> So I''m not considering one of LSI''s HBA''s - what do list members think about this device: <http://www.provantage.com/lsi-logic-lsi00117~7LSIG03X.htm> On Thu, Mar 12, 2009 at 2:18 AM, Nathan Kroenert <Nathan.Kroenert at sun.com> wrote:> definitely time to bust out some mdb -K or boot -k and see what it''s moaning > about. > > I did not see the screenshot earlier... sorry about that. > > Nathan. > > Blake wrote: >> >> I start the cp, and then, with prstat -a, watch the cpu load for the >> cp process climb to 25% on a 4-core machine. >> >> Load, measured for example with ''uptime'', climbs steadily until the >> reboot. >> >> Note that the machine does not dump properly, panic or hang - rather, >> it reboots. >> >> I attached a screenshot earlier in this thread of the little bit of >> error message I could see on the console. ?The machine is trying to >> dump to the dump zvol, but fails to do so. ?Only sometimes do I see an >> error on the machine''s local console - mos times, it simply reboots. >> >> >> >> On Thu, Mar 12, 2009 at 1:55 AM, Nathan Kroenert >> <Nathan.Kroenert at sun.com> wrote: >>> >>> Hm - >>> >>> Crashes, or hangs? Moreover - how do you know a CPU is pegged? >>> >>> Seems like we could do a little more discovery on what the actual problem >>> here is, as I can read it about 4 different ways. >>> >>> By this last piece of information, I''m guessing the system does not >>> crash, >>> but goes really really slow?? >>> >>> Crash == panic == we see stack dump on console and try to take a dump >>> hang == nothing works == no response -> might be worth looking at mdb -K >>> ? ? ? or booting with a -k on the boot line. >>> >>> So - are we crashing, hanging, or something different? >>> >>> It might simply be that you are eating up all your memory, and your >>> physical >>> backing storage is taking a while to catch up....? >>> >>> Nathan. >>> >>> Blake wrote: >>>> >>>> My dump device is already on a different controller - the motherboards >>>> built-in nVidia SATA controller. >>>> >>>> The raidz2 vdev is the one I''m having trouble with (copying the same >>>> files to the mirrored rpool on the nVidia controller work nicely). ?I >>>> do notice that, when using cp to copy the files to the raidz2 pool, >>>> load on the machine climbs steadily until the crash, and one proc core >>>> pegs at 100%. >>>> >>>> Frustrating, yes. >>>> >>>> On Thu, Mar 12, 2009 at 12:31 AM, Maidak Alexander J >>>> <MaidakAlexanderJ at johndeere.com> wrote: >>>>> >>>>> If you''re having issues with a disk contoller or disk IO driver its >>>>> highly likely that a savecore to disk after the panic will fail. ?I''m >>>>> not >>>>> sure how to work around this, maybe a dedicated dump device not on a >>>>> controller that uses a different driver then the one that you''re having >>>>> issues with? >>>>> >>>>> -----Original Message----- >>>>> From: zfs-discuss-bounces at opensolaris.org >>>>> [mailto:zfs-discuss-bounces at opensolaris.org] On Behalf Of Blake >>>>> Sent: Wednesday, March 11, 2009 4:45 PM >>>>> To: Richard Elling >>>>> Cc: Marc Bevand; zfs-discuss at opensolaris.org >>>>> Subject: Re: [zfs-discuss] reboot when copying large amounts of data >>>>> >>>>> I guess I didn''t make it clear that I had already tried using savecore >>>>> to >>>>> retrieve the core from the dump device. >>>>> >>>>> I added a larger zvol for dump, to make sure that I wasn''t running out >>>>> of >>>>> space on the dump device: >>>>> >>>>> root at host:~# dumpadm >>>>> ? ?Dump content: kernel pages >>>>> ? ? Dump device: /dev/zvol/dsk/rpool/bigdump (dedicated) Savecore >>>>> directory: /var/crash/host >>>>> ?Savecore enabled: yes >>>>> >>>>> I was using the -L option only to try to get some idea of why the >>>>> system >>>>> load was climbing to 1 during a simple file copy. >>>>> >>>>> >>>>> >>>>> On Wed, Mar 11, 2009 at 4:58 PM, Richard Elling >>>>> <richard.elling at gmail.com> wrote: >>>>>> >>>>>> Blake wrote: >>>>>>> >>>>>>> I''m attaching a screenshot of the console just before reboot. ?The >>>>>>> dump doesn''t seem to be working, or savecore isn''t working. >>>>>>> >>>>>>> On Wed, Mar 11, 2009 at 11:33 AM, Blake <blake.irvin at gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> I''m working on testing this some more by doing a savecore -L right >>>>>>>> after I start the copy. >>>>>>>> >>>>>>>> >>>>>> savecore -L is not what you want. >>>>>> >>>>>> By default, for OpenSolaris, savecore on boot is disabled. ?But the >>>>>> core will have been dumped into the dump slice, which is not used for >>>>>> swap. >>>>>> So you should be able to run savecore at a later time to collect the >>>>>> core from the last dump. >>>>>> -- richard >>>>>> >>>>>> >>>>> _______________________________________________ >>>>> zfs-discuss mailing list >>>>> zfs-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>>>> >>>> _______________________________________________ >>>> zfs-discuss mailing list >>>> zfs-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >>> >>> -- >>> ////////////////////////////////////////////////////////////////// >>> // Nathan Kroenert ? ? ? ? ? ? ?nathan.kroenert at sun.com ? ? ? ? // >>> // Systems Engineer ? ? ? ? ? ? Phone: ?+61 3 9869-6255 ? ? ? ? // >>> // Sun Microsystems ? ? ? ? ? ? Fax: ? ?+61 3 9869-6288 ? ? ? ? // >>> // Level 7, 476 St. Kilda Road ?Mobile: 0419 305 456 ? ? ? ? ? ?// >>> // Melbourne 3004 ? Victoria ? ?Australia ? ? ? ? ? ? ? ? ? ? ? // >>> ////////////////////////////////////////////////////////////////// >>> > > -- > ////////////////////////////////////////////////////////////////// > // Nathan Kroenert ? ? ? ? ? ? ?nathan.kroenert at sun.com ? ? ? ? // > // Systems Engineer ? ? ? ? ? ? Phone: ?+61 3 9869-6255 ? ? ? ? // > // Sun Microsystems ? ? ? ? ? ? Fax: ? ?+61 3 9869-6288 ? ? ? ? // > // Level 7, 476 St. Kilda Road ?Mobile: 0419 305 456 ? ? ? ? ? ?// > // Melbourne 3004 ? Victoria ? ?Australia ? ? ? ? ? ? ? ? ? ? ? // > ////////////////////////////////////////////////////////////////// >
On Thu, Mar 12, 2009 at 2:22 PM, Blake <blake.irvin at gmail.com> wrote:> I''ve managed to get the data transfer to work by rearranging my disks > so that all of them sit on the integrated SATA controller. > > So, I feel pretty certain that this is either an issue with the > Supermicro aoc-sat2-mv8 card, or with PCI-X on the motherboard (though > I would think that the integrated SATA would also be using the PCI > bus?). > > The motherboard, for those interested, is an HD8ME-2 (not, I now find > after buying this box from Silicon Mechanics, a board that''s on the > Solaris HCL...) > > <http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/h8dme-2.cfm > > > > So I''m not considering one of LSI''s HBA''s - what do list members think > about this device: > > <http://www.provantage.com/lsi-logic-lsi00117~7LSIG03X.htm<http://www.provantage.com/lsi-logic-lsi00117%7E7LSIG03X.htm> > > > >I believe the MCP55''s SATA controllers are actually PCI-E based. --Tim -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090312/399456c0/attachment.html>
Tim wrote:> > > On Thu, Mar 12, 2009 at 2:22 PM, Blake <blake.irvin at gmail.com > <mailto:blake.irvin at gmail.com>> wrote: > > I''ve managed to get the data transfer to work by rearranging my disks > so that all of them sit on the integrated SATA controller. > > So, I feel pretty certain that this is either an issue with the > Supermicro aoc-sat2-mv8 card, or with PCI-X on the motherboard (though > I would think that the integrated SATA would also be using the PCI > bus?). > > The motherboard, for those interested, is an HD8ME-2 (not, I now find > after buying this box from Silicon Mechanics, a board that''s on the > Solaris HCL...) > > <http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/h8dme-2.cfm> > > So I''m not considering one of LSI''s HBA''s - what do list members think > about this device: > > <http://www.provantage.com/lsi-logic-lsi00117~7LSIG03X.htm > <http://www.provantage.com/lsi-logic-lsi00117%7E7LSIG03X.htm>> > > > > I believe the MCP55''s SATA controllers are actually PCI-E based.I use Tyan 2927 motherboards. They have on-board nVidia MCP55 chipsets, which is the same chipset at the X4500 (IIRC). I wouldn''t trust the MCP55 chipset in OpenSolaris. I had random disk hangs even while the machine was mostly idle. In Feb 2008 I bought AOC-SAT2-MV8 cards and moved all my drives to these add-in cards. I haven''t had any issues with drive hanging since. There does not seem to be any problems with the SAT2-MV8 under heavy load in my servers from what I''ve seen. When the SuperMicro AOC-USAS-L8i came out later last year, I started using them instead. They work better than the SAT2-MV8s. This card needs a 3U or bigger case: http://www.supermicro.com/products/accessories/addon/AOC-USAS-L8i.cfm This is the low profile card that will fit in a 2U: http://www.supermicro.com/products/accessories/addon/AOC-USASLP-L8i.cfm They both work in normal PCI-E slots on my Tyan 2927 mobos. Finding good non-Sun hardware that works very well under OpenSolaris is frustrating to say the least. Good luck. -- Dave
Miles Nordin
2009-Mar-12 22:30 UTC
[zfs-discuss] reboot when copying large amounts of data
>>>>> "b" == Blake <blake.irvin at gmail.com> writes:b> http://www.provantage.com/lsi-logic-lsi00117~7LSIG03X.htm I''m having trouble matching up chips, cards, drivers, platforms, and modes with the LSI stuff. The more I look at it the mroe confused I get. Platforms: x86 SPARC Drivers: mpt mega_sas mfi Chips: 1068 (SAS, PCI-X) 1068E (SAS, PCIe) 1078 ??? -- from supermicro, seems to be SAS, PCIe, with support for 256 - 512MB RAM instead of the 16 - 32MB RAM on the others 1030 (parallel scsi) Cards: LSI cards http://www.lsi.com/storage_home/products_home/host_bus_adapters/sas_hbas/index.html I love the way they use the numbers 3800 and 3080, so you are constantly transposing them thus leaving google littered with all this confusingly wrong information. LSISAS3800X (PCI-X, external ports) LSISAS3080X-R (PCI-X, internal ports) LSISAS3801X (PCI-X, external ports) LSISAS3801E (PCIe, external ports) LSISAS3081E-R (PCIe, internal ports) I would have thought -R meant ``suports RAID'''' but all I can really glean through the foggy marketing-glass behind which all the information is hidden, is -R means ``all the ports are internal''''. Supermicro cards http://www.supermicro.com/products/accessories/index.cfm wow, this is even more of a mess. These are all UIO cards so I assume they have the PCIe bracket on backwards AOC-USAS-L4i (PCIe, 4 internal 4 external) AOC-USAS-L8i, AOC-USASLP-L8i (PCIe, internal ports) based on 1068E sounds similar to LSISAS3081E. Is that also 1068E? supports RAID0, RAID1, RAID10 AOC-USAS-L4iR identical to the above, but ``includes iButton'''' which is an old type of smartcard-like device with sometimes crypto and javacard support. apparently some kind of license key to unlock RAID5? no L8iR exists though, only L4iR. I have the L8i, and it does have an iButton socket with no button in it. AOC-USAS-H4iR AOC-USAS-H8iR, AOC-USASLP-H8iR (PCIe, internal ports) based on 1078 low-profile version has more memory than fullsize version?! but here is the most fun thing about the supermicro cards. All cards have one driver *EXCEPT* the L8i, which has three drivers for three modes: IT, IR, and SR. When I google for this I find notes on some of their integrated motherboards like: * The onboard LSI 1068E supported SR and IT mode but not IR mode. I also found this: * SR = Software RAID IT = Integrate. Target mode. IR mode is not supported. but no idea what the three modes are. searching for SAS SR IT IR doesn''t work either, so it''s not some SAS thing. What *is* it? also there seem to be two different kinds of quad-SATA connector on these SAS cards so there are two different kinds of octopus cable. Questions: * which chips are used by each of the LSI boards? I can guess, but in particular LSISAS3800X and LSISAS3801X seem to be different chips, while from the list of chips I''d have no choice but to guess they are both 1068. * which drivers work on x86 and which SPARC? I know some LSI cards work in SPARC but maybe not all---do the drivers support the same set of cards on both platforms? Or will normal cards not work in SPARC for lack of Forth firmware to perform some LSI-proprietary ``initialization'''' ritual? * which chips go with which drivers? Is it even that simple---will adding an iButton RAID5 license to a SuperMicro board make the same card change from mega_sas to mpt attachment, or something similar? For example there is a bug here about a 1068E card which doesn''t work, even though most 1068E cards do work: http://bugs.opensolaris.org/view_bug.do?bug_id=6736187 Maybe the Solaris driver needs IR mode and won''t work with the onboard supermicro chip which supports only ``software raid'''' whatever that means, which is maybe denoted by SR? What does the iButton unlock, then, features of IR mode which are abstracted from the OS driver? * What are SR, IT, and IR mode? Which modes do the Solaris drivers use, or does it matter? * Has someone found the tool mentioned here by some above-the-table means, or only by request from LSI?: http://www.opensolaris.org/jive/message.jspa?messageID=184811#184811 The mention that a SPARC version of the tool exists is encouraging. The procedure to clear persistent mappings through the BIOS obviously won''t work on SPARC. Here are the notes I have so far: -----8<-----> The driver for LSI''s MegaRAID SAS card is "mega_sas" which > was integrated into snv_88. It''s planned for backporting to > a Solaris 10 update.There is also a BSD-licensed driver for that hardware, called "mfi". It''s available from http://www.itee.uq.edu.au/~dlg/mfi> a scsi_vhci > sort of driver for the LSI card in the Ultra {20,25}Well yes, that''s mpt(7d) as delivered into NV build 63, and backported to Solaris 10 Update 5, found in patch 125081-14 and 125082-14. We''ve got support for both SAS (1064/E, 1068/E and 1078) and Parallel SCSI (1030) chips from LSI in that driver. SATA disks will always show up when attached to a SAS HBA, because that''s one of the requirements of the SAS specification.> LSISAS3801EI think you might actually be referring to the LSI SAS3801-R You''re correct in that it''s not using mpt, but mega_sas or mfi. But it''s not a SATA framework driver. -----8<----- which doesn''t seem to be internally consistent. note that SAS3801-R is not a product that exists on LSI''s site. There is LSISAS3801X (without the -R), or LSISAS3081E-R, LSISAS3080X-R (with -R). note also that 1068, 1068E, and 1078 in the middle statement suggests mpt supports every LSI card we are discussing (probably---LSI doesn''t say what chips are on each card), while the last statement contradicts it by mentioning a nonexistant card of the same era we are discussing which doesn''t work with mpt driver. -- READ CAREFULLY. By reading this fortune, you agree, on behalf of your employer, to release me from all obligations and waivers arising from any and all NON-NEGOTIATED agreements, licenses, terms-of-service, shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure, non-compete and acceptable use policies ("BOGUS AGREEMENTS") that I have entered into with your employer, its partners, licensors, agents and assigns, in perpetuity, without prejudice to my ongoing rights and privileges. You further represent that you have the authority to release me from any BOGUS AGREEMENTS on behalf of your employer. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 304 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090312/c0bb2178/attachment.bin>
Nathan Kroenert
2009-Mar-12 23:18 UTC
[zfs-discuss] reboot when copying large amounts of data
For what it''s worth, I have been running Nevada (so, same kernel as opensolaris) for ages (at least 18 months) on a Gigabyte board with the MCP55 chipset and it''s been flawless. I liked it so much, I bought it''s newer brother, based on the nvidia 750SLI chipset... M750SLI-DS4 Cheers! Nathan. On 13/03/09 09:21 AM, Dave wrote:> > > Tim wrote: >> >> >> On Thu, Mar 12, 2009 at 2:22 PM, Blake <blake.irvin at gmail.com >> <mailto:blake.irvin at gmail.com>> wrote: >> >> I''ve managed to get the data transfer to work by rearranging my disks >> so that all of them sit on the integrated SATA controller. >> >> So, I feel pretty certain that this is either an issue with the >> Supermicro aoc-sat2-mv8 card, or with PCI-X on the motherboard >> (though >> I would think that the integrated SATA would also be using the PCI >> bus?). >> >> The motherboard, for those interested, is an HD8ME-2 (not, I now find >> after buying this box from Silicon Mechanics, a board that''s on the >> Solaris HCL...) >> >> >> <http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/h8dme-2.cfm> >> >> >> So I''m not considering one of LSI''s HBA''s - what do list members >> think >> about this device: >> >> <http://www.provantage.com/lsi-logic-lsi00117~7LSIG03X.htm >> <http://www.provantage.com/lsi-logic-lsi00117%7E7LSIG03X.htm>> >> >> >> >> I believe the MCP55''s SATA controllers are actually PCI-E based. > > I use Tyan 2927 motherboards. They have on-board nVidia MCP55 chipsets, > which is the same chipset at the X4500 (IIRC). I wouldn''t trust the > MCP55 chipset in OpenSolaris. I had random disk hangs even while the > machine was mostly idle. > > In Feb 2008 I bought AOC-SAT2-MV8 cards and moved all my drives to these > add-in cards. I haven''t had any issues with drive hanging since. There > does not seem to be any problems with the SAT2-MV8 under heavy load in > my servers from what I''ve seen. > > When the SuperMicro AOC-USAS-L8i came out later last year, I started > using them instead. They work better than the SAT2-MV8s. > > This card needs a 3U or bigger case: > http://www.supermicro.com/products/accessories/addon/AOC-USAS-L8i.cfm > > This is the low profile card that will fit in a 2U: > http://www.supermicro.com/products/accessories/addon/AOC-USASLP-L8i.cfm > > They both work in normal PCI-E slots on my Tyan 2927 mobos. > > Finding good non-Sun hardware that works very well under OpenSolaris is > frustrating to say the least. Good luck. > > -- > Dave > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss-- ////////////////////////////////////////////////////////////////// // Nathan Kroenert nathan.kroenert at sun.com // // Senior Systems Engineer Phone: +61 3 9869 6255 // // Global Systems Engineering Fax: +61 3 9869 6288 // // Level 7, 476 St. Kilda Road // // Melbourne 3004 Victoria Australia // //////////////////////////////////////////////////////////////////
Will Murnane
2009-Mar-12 23:42 UTC
[zfs-discuss] reboot when copying large amounts of data
On Thu, Mar 12, 2009 at 18:30, Miles Nordin <carton at ivy.net> wrote:> ?I love the way they use the numbers 3800 and 3080, so you are > ?constantly transposing them thus leaving google littered with all > ?this confusingly wrong information.Think of the middle two digits as (number of external ports, number of internal ports). For example, I have a 3442E-R which has 4 internal and 4 external ports, the 3800 has 8 external ports and 0 internal, and so forth. One place this breaks down is with cards like the 8888; it has a total of 8 ports, any group of 4 of which can be mapped to internal or external ports.> ? AOC-USAS-L4iR > ? ? ? identical to the above, but ``includes iButton'''' > ? ? ? which is an old type of smartcard-like device with > ? ? ? sometimes crypto and javacard support. > ? ? ? apparently some kind of license key to > ? ? ? unlock RAID5? ?no L8iR exists though, only L4iR. > ? ? ? I have the L8i, and it does have an iButton socket > ? ? ? with no button in it.I think the iButton is just used as an unlock code for the builtin RAID 5 functionality. Nothing the end user cares about, unless they want RAID and have to spend the extra money.> ? ? * SR = Software RAID IT = Integrate. Target mode. IR mode is not supported.Integrated target mode lets you export some storage attached to the host system (through another adapter, presumably) as a storage device. IR mode is almost certainly Internal RAID, which that card doesn''t have support for.> also there seem to be two different kinds of quad-SATA connector on > these SAS cards so there are two different kinds of octopus cable.Yes---SFF-8484 and SFF-8087 are the key words.> SATA disks will always show up when attached to a SAS HBA, > because that''s one of the requirements of the SAS specification.I''m not sure what you mean by this. SAS controllers can control SATA disks, and interact with them. They don''t just "show up"; they''re first-class citizens. Will
Miles Nordin
2009-Mar-13 02:24 UTC
[zfs-discuss] reboot when copying large amounts of data
>>>>> "wm" == Will Murnane <will.murnane at gmail.com> writes:>> ? ? * SR = Software RAID IT = Integrate. Target mode. IR mode >> is not supported. wm> Integrated target mode lets you export some storage attached wm> to the host system (through another adapter, presumably) as a wm> storage device. IR mode is almost certainly Internal RAID, wm> which that card doesn''t have support for. no, the supermicro page for AOC-USAS-L8i does claim support for all three, and supermicro has an ``IR driver'''' available for download for Linux and Windows, or at least a link to one. I''m trying to figure out what''s involved in determining and switching modes, why you''d want to switch them, what cards support which modes, which solaris drivers support which modes, u.s.w. The answer may be very simple, like ``the driver supports only IR. Most cards support IR, and cards that don''t support IR won''t work. IR can run in single-LUN mode. Some IR cards support RAID5, others support only RAID 0, 1, 10.'''' Or it could be ``the driver supports only SR. The driver is what determines the mode, and it does this by loading firmware into the card, and the first step in initializing the card is always for the driver to load in a firmware blob. All currently-produced cards support SR.'''' so...actually, now that I say it, I guess the answer cannot be very simple. It''s going to have to be a little complicated. Anyway, I can guess, too. I was hoping someone would know for sure off-hand. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 304 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090312/9264a7b7/attachment.bin>
James C. McPherson
2009-Mar-13 02:42 UTC
[zfs-discuss] reboot when copying large amounts of data
On Thu, 12 Mar 2009 22:24:12 -0400 Miles Nordin <carton at Ivy.NET> wrote:> >>>>> "wm" == Will Murnane <will.murnane at gmail.com> writes: > > >> ? ? * SR = Software RAID IT = Integrate. Target mode. IR mode > >> is not supported. > wm> Integrated target mode lets you export some storage attached > wm> to the host system (through another adapter, presumably) as a > wm> storage device. IR mode is almost certainly Internal RAID, > wm> which that card doesn''t have support for. > > no, the supermicro page for AOC-USAS-L8i does claim support for all > three, and supermicro has an ``IR driver'''' available for download for > Linux and Windows, or at least a link to one. > > I''m trying to figure out what''s involved in determining and switching > modes, why you''d want to switch them, what cards support which modes, > which solaris drivers support which modes, u.s.w. > > The answer may be very simple, like ``the driver supports only IR. > Most cards support IR, and cards that don''t support IR won''t work. IR > can run in single-LUN mode. Some IR cards support RAID5, others > support only RAID 0, 1, 10.'''' Or it could be ``the driver supports > only SR. The driver is what determines the mode, and it does this by > loading firmware into the card, and the first step in initializing the > card is always for the driver to load in a firmware blob. All > currently-produced cards support SR.'''' so...actually, now that I say > it, I guess the answer cannot be very simple. It''s going to have to > be a little complicated. > Anyway, I can guess, too. I was hoping someone would know for sure > off-hand.Hi Miles, the mpt(7D) driver supports that card. mpt(7D) supports both IT and IR firmware variants. You can find out the specifics for what RAID volume levels are supported by reading the raidctl(1M) manpage. I don''t think you can switch between IT and IR firmware, but not having needed to know this before, I haven''t tried it. James C. McPherson -- Senior Kernel Software Engineer, Solaris Sun Microsystems http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog
This is really great information, though most of the controllers mentioned aren''t on the OpenSolaris HCL. Seems like that should be corrected :) My thanks to the community for their support. On Mar 12, 2009, at 10:42 PM, "James C. McPherson" <James.McPherson at Sun.COM > wrote:> On Thu, 12 Mar 2009 22:24:12 -0400 > Miles Nordin <carton at Ivy.NET> wrote: > >>>>>>> "wm" == Will Murnane <will.murnane at gmail.com> writes: >> >>>> * SR = Software RAID IT = Integrate. Target mode. IR mode >>>> is not supported. >> wm> Integrated target mode lets you export some storage attached >> wm> to the host system (through another adapter, presumably) as a >> wm> storage device. IR mode is almost certainly Internal RAID, >> wm> which that card doesn''t have support for. >> >> no, the supermicro page for AOC-USAS-L8i does claim support for all >> three, and supermicro has an ``IR driver'''' available for download for >> Linux and Windows, or at least a link to one. >> >> I''m trying to figure out what''s involved in determining and switching >> modes, why you''d want to switch them, what cards support which modes, >> which solaris drivers support which modes, u.s.w. >> >> The answer may be very simple, like ``the driver supports only IR. >> Most cards support IR, and cards that don''t support IR won''t work. >> IR >> can run in single-LUN mode. Some IR cards support RAID5, others >> support only RAID 0, 1, 10.'''' Or it could be ``the driver supports >> only SR. The driver is what determines the mode, and it does this by >> loading firmware into the card, and the first step in initializing >> the >> card is always for the driver to load in a firmware blob. All >> currently-produced cards support SR.'''' so...actually, now that I say >> it, I guess the answer cannot be very simple. It''s going to have to >> be a little complicated. >> Anyway, I can guess, too. I was hoping someone would know for sure >> off-hand. > > > Hi Miles, > the mpt(7D) driver supports that card. mpt(7D) supports both > IT and IR firmware variants. You can find out the specifics > for what RAID volume levels are supported by reading the > raidctl(1M) manpage. I don''t think you can switch between IT > and IR firmware, but not having needed to know this before, > I haven''t tried it. > > > James C. McPherson > -- > Senior Kernel Software Engineer, Solaris > Sun Microsystems > http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Hi folks, I was trying to load a large file in /tmp so that a process that parses it wouldn''t be limited by a disk throughput bottleneck. My rig here has only 12GB of RAM and the file I copied is about 12GB as well. Before the copied finished, my system restarted. I''m pretty up to date, the system has b129 running. Googling about this I stumbled upon this thread dating back to March ... http://mail.opensolaris.org/pipermail/zfs-discuss/2009-March/027264.html Since this seems to be a known issue, and pretty serious in my book, is there a fix pending or is that not investigated ? thanks in advance and happy holidays to all -=arnaud=-
arnaud wrote:> Hi folks, > I was trying to load a large file in /tmp so that a process that > parses it wouldn''t be limited by a disk throughput bottleneck. My rig > here has only 12GB of RAM and the file I copied is about 12GB as well. > Before the copied finished, my system restarted. > I''m pretty up to date, the system has b129 running. >12GB of ECC RAM? This kind of spontaneous reboot with crash dumps enabled (do you have them enabled?) tend to be hardware related. The most common bit of bad hardware is RAM... -- Ian.