So looking at the README for patch 14144[45]-09, there are ton of ZFS fixes and feature adds. The big features are already described in the update 8 release docs, but would anyone in-the-know care to comment or point out any interesting CR fixes that might be substantial in the areas of stability or performance? Thanks for what looks like a well-loaded KJP. /dale
Dale Ghent wrote:> > So looking at the README for patch 14144[45]-09, there are ton of ZFS > fixes and feature adds. > > The big features are already described in the update 8 release docs, > but would anyone in-the-know care to comment or point out any > interesting CR fixes that might be substantial in the areas of > stability or performance?A couple of my CRs (a panic and a hang) are fixed in there, but LU appears to be FUBAR for systems with ZFS root and zones, so I can''t run any tests.... -- Ian.
On Thu, Oct 15, 2009 at 2:57 AM, Ian Collins <ian at ianshome.com> wrote:> Dale Ghent wrote: >> >> So looking at the README for patch 14144[45]-09, there are ton of ZFS >> fixes and feature adds. >> >> The big features are already described in the update 8 release docs, but >> would anyone in-the-know care to comment or point out any interesting CR >> fixes that might be substantial in the areas of stability or performance? > > A couple of my CRs (a panic and a hang) are fixed in there, but LU appears > to be FUBAR for systems with ZFS root and zones, so I can''t run any > tests....Having tried to install about 5 patches on a system with ZFS root + sparse zones (plus a delegated dataset), FUBAR is putting it mildly.. :) I found upgrade on attach worked much better in that instance (just meant I could only snapshot, not create a new BE). But hopefully I can get ahold of a box for more testing to get it to actually work.> > -- > Ian. > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >
Jason King wrote:> On Thu, Oct 15, 2009 at 2:57 AM, Ian Collins <ian at ianshome.com> wrote: >> Dale Ghent wrote: >>> So looking at the README for patch 14144[45]-09, there are ton of ZFS >>> fixes and feature adds. >>> >>> The big features are already described in the update 8 release docs, but >>> would anyone in-the-know care to comment or point out any interesting CR >>> fixes that might be substantial in the areas of stability or performance? >> A couple of my CRs (a panic and a hang) are fixed in there, but LU appears >> to be FUBAR for systems with ZFS root and zones, so I can''t run any >> tests.... > > Having tried to install about 5 patches on a system with ZFS root + > sparse zones (plus a delegated dataset), FUBAR is putting it mildly.. > :)if you have a separate /var dataset on zfs root then Lu in update 8 ( or using latest 121430-42/121431-43 ) is broke. this is covered in CR 6884728 So you do need to apply 119255-70 before patching zfs root on x86 as well, to avoid another issue. what was the error/problem you ran into, output from patchadd + logs from /var/sadm/patch/*/log for failed patch, or /var/tmp/<patchid>.log.$$ if they exist, plus some data on the setup, ie zfs list and zonecfg to give an idea. Enda> > I found upgrade on attach worked much better in that instance (just > meant I could only snapshot, not create a new BE). But hopefully I > can get ahold of a box for more testing to get it to actually work. > >> -- >> Ian. >> >> _______________________________________________ >> 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-- Enda O''Connor x19781 Software Product Engineering Patch System Test : Ireland : x19781/353-1-8199718
On Thu, Oct 15, 2009 at 9:25 AM, Enda O''Connor <Enda.Oconnor at sun.com> wrote:> > > Jason King wrote: >> >> On Thu, Oct 15, 2009 at 2:57 AM, Ian Collins <ian at ianshome.com> wrote: >>> >>> Dale Ghent wrote: >>>> >>>> So looking at the README for patch 14144[45]-09, there are ton of ZFS >>>> fixes and feature adds. >>>> >>>> The big features are already described in the update 8 release docs, but >>>> would anyone in-the-know care to comment or point out any interesting CR >>>> fixes that might be substantial in the areas of stability or >>>> performance? >>> >>> A couple of my CRs (a panic and a hang) are fixed in there, but LU >>> appears >>> to be FUBAR for systems with ZFS root and zones, so I can''t run any >>> tests.... >> >> Having tried to install about 5 patches on a system with ZFS root + >> sparse zones (plus a delegated dataset), FUBAR is putting it mildly.. >> :) > > if you have a separate /var dataset on zfs root then Lu in update 8 ( or > using latest 121430-42/121431-43 ) is broke. > this is covered in CR 6884728 > > So you do need to apply 119255-70 before patching zfs root on x86 as well, > to avoid another issue. > > what was the error/problem you ran into, output from patchadd + logs from > /var/sadm/patch/*/log for failed patch, or /var/tmp/<patchid>.log.$$ if they > exist, plus some data on the setup, ie zfs list and zonecfg to give an idea.My problem was creating the new BE to patch -- lucreate worked, but lumount was a disaster and left things so horribly messed up that even after a luumount and ludelete of the new BE, a reboot was required just to make the system sane (thankfully this was all in a maint window anyway). But it caused a bunch of stale mnttab entries that wouldn''t go away as well as a bunch of ''already mounted'' errors when you tried to do anything with the filesystems. I punted, rebooted to clear things up, did a zfs snapshot of everything, then patched the lived system (since I had a boot server I could use to mount the pool and rollback if needed).> > > Enda > > > >> >> I found upgrade on attach worked much better in that instance (just >> meant I could only snapshot, not create a new BE). ?But hopefully I >> can get ahold of a box for more testing to get it to actually work. >> >>> -- >>> Ian. >>> >>> _______________________________________________ >>> 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 > > -- > Enda O''Connor x19781 ?Software Product Engineering > Patch System Test : Ireland : x19781/353-1-8199718 >
Enda O''Connor wrote:> > > Jason King wrote: >> On Thu, Oct 15, 2009 at 2:57 AM, Ian Collins <ian at ianshome.com> wrote: >>> Dale Ghent wrote: >>>> So looking at the README for patch 14144[45]-09, there are ton of ZFS >>>> fixes and feature adds. >>>> >>>> The big features are already described in the update 8 release >>>> docs, but >>>> would anyone in-the-know care to comment or point out any >>>> interesting CR >>>> fixes that might be substantial in the areas of stability or >>>> performance? >>> A couple of my CRs (a panic and a hang) are fixed in there, but LU >>> appears >>> to be FUBAR for systems with ZFS root and zones, so I can''t run any >>> tests.... >> >> Having tried to install about 5 patches on a system with ZFS root + >> sparse zones (plus a delegated dataset), FUBAR is putting it mildly.. >> :) > > if you have a separate /var dataset on zfs root then Lu in update 8 ( > or using latest 121430-42/121431-43 ) is broke. > this is covered in CR 6884728 >It appears to be broken if the system has zones as well. From an luucreate log: + /sbin/zfs set zpdata:rbe=10u8 rpool/ROOT/10u8/zoneRoot/common-10u8 + /sbin/zfs set zpdata:zn=common rpool/ROOT/10u8/zoneRoot/common-10u8 + echo /zoneRoot/common + sed s:^//:/: pbe_rawzp=/zoneRoot/common + zfs get -Ho value mountpoint rpool/ROOT/10u7ZFSa/zoneRoot/common mount_prop=/zoneRoot/common + [ /zoneRoot/common = legacy ] + /sbin/zfs get -Ho value mountpoint rpool/ROOT/10u8/zoneRoot/common-10u8 newpath=legacy newrawpath=legacy + /sbin/zfs mount rpool/ROOT/10u8/zoneRoot/common-10u8 cannot mount ''rpool/ROOT/10u8/zoneRoot/common-10u8'': legacy mountpoint use mount(1M) to mount this filesystem + [ 1 -ne 0 ] + gettext Failed to mount dataset <%s> + /etc/lib/lu/luprintf -Eelp2 Failed to mount dataset <%s> rpool/ROOT/10u8/zoneRoot/common-10u8 luclonefs: ERROR: Failed to mount dataset <rpool/ROOT/10u8/zoneRoot/common-10u8> + [ -n rpool/ROOT/10u7ZFSa/zoneRoot/common ] + /sbin/zfs set canmount=noauto rpool/ROOT/10u8/zoneRoot/common-10u8 + zonecfg -R /.alt.tmp.b-Mdh.mnt -z common set -F zonepath=legacy -- Ian.
On Fri, 16 Oct 2009 07:06:23 +1300 Ian Collins <ian at ianshome.com> wrote:> + [ /zoneRoot/common = legacy ]ABSURT. I''m sure the zone has it''s own mountpoint set. And now LU presumes all of a sudden a legacy mp? (It has even set it in /etc/vfstab in the new ABE).> cannot mount ''rpool/ROOT/10u8/zoneRoot/common-10u8'': legacy mountpoint > use mount(1M) to mount this filesystemThat''s how LU scr*&s things up. -- Dick Hoogendijk -- PGP/GnuPG key: 01D2433D + http://nagual.nl/ | SunOS 10u8 10/09 | OpenSolaris 2010.02 b123 + All that''s really worth doing is what we do for others (Lewis Carrol)
On Thu, 15 Oct 2009, Enda O''Connor wrote:> if you have a separate /var dataset on zfs root then Lu in update 8 ( or > using latest 121430-42/121431-43 ) is broke. this is covered in CR > 6884728I don''t see that bugid available in bugs.opensolaris.org, is there any place I can find more details on that? -- Paul B. Henson | (909) 979-6361 | http://www.csupomona.edu/~henson/ Operating Systems and Network Analyst | henson at csupomona.edu California State Polytechnic University | Pomona CA 91768