Hi .. Having read the man page, there is alot in there that makes me wonder whether going with a HP Smart Array P600 is a wise idea ... "The Compaq ciss adapters require faked responses to get reasonable behavior out of them. In addition, the ciss command set is by no means adequate to support the functionality of a RAID controller, and thus the supported Compaq adapters utilize portions of the control protocol from earlier" I'm specifically looking at the Proliant DL360, which has this card ... can you provide any comments, or insight, concerning what the man page states? Should I shy away from this controller? :( Thanks ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
> I'm specifically looking at the Proliant DL360, which has this card ... > can you provide any comments, or insight, concerning what the man page > states? Should I shy away from this controller? :(I use SMART controllers driven by the ciss driver, as well as earlier ones driven using ida, under 5.4 with no problem and have used them on previous FreeBSD's for a long time. They work fine for me and give adequate performance. I dont see any reason why they wont work nicely under 6.0 aas well (indeed I will be upgrading a machine with such a controller to 6.0 over the next few days) I wouldn't worry about it if I were you -pcf.
On Wed, Nov 23, 2005 at 03:37:23AM -0400, Marc G. Fournier wrote:> > Hi .. > > Having read the man page, there is alot in there that makes me wonder > whether going with a HP Smart Array P600 is a wise idea ... > > "The Compaq ciss adapters require faked responses to get reasonable > behavior out of them. In addition, the ciss command set is by no means > adequate to support the functionality of a RAID controller, and thus the > supported Compaq adapters utilize portions of the control protocol from > earlier" > > I'm specifically looking at the Proliant DL360, which has this card ... > can you provide any comments, or insight, concerning what the man page > states? Should I shy away from this controller? :(I have so far not experienced any problems with ciss-controllers on Proliant G2, G3, G4, especially not with DL360 and DL360 (or MPxxx either) ciss0: <Compaq Smart Array 5i> port 0x2800-0x28ff mem 0xf5df0000-0xf5df3fff,0xf5f80000-0xf5fbffff irq 31 at device 4.0 on pci0 ciss0: [GIANT-LOCKED] ciss0@pci0:4:0: class=0x010400 card=0x40800e11 chip=0xb1780e11 rev=0x01 hdr=0x00 vendor = 'Compaq Computer Corp (Now owned by Hewlett-Packard)' device = 'CISSB SMART2 Array Controller' class = mass storage subclass = RAID --- ciss0: <HP Smart Array 641> port 0x3000-0x30ff mem 0xf7f80000-0xf7fbffff,0xf7ff0000-0xf7ff1fff irq 24 at device 2.0 on pci2 ciss0: [GIANT-LOCKED] ciss0@pci2:2:0: class=0x010400 card=0x409a0e11 chip=0x00460e11 rev=0x01 hdr=0x00 vendor = 'Compaq Computer Corp (Now owned by Hewlett-Packard)' device = 'Smart Array 64xx Controller' class = mass storage subclass = RAID I especially like that rebuilding of raid 1 and 5 is done automatically. You'll get entries via syslog about failing drives, and just removing that drive, inserting a spare one, will start rebuilding which is also syslogged. Very nice. No Problems so far, and no system failure. Regular backkups still recommended, though ;-) Regards, Holger Kipp
Marc G. Fournier wrote:> > Hi ..Hi,> > Having read the man page, there is alot in there that makes me wonder > whether going with a HP Smart Array P600 is a wise idea ... > > "The Compaq ciss adapters require faked responses to get reasonable > behavior out of them. In addition, the ciss command set is by no means > adequate to support the functionality of a RAID controller, and thus the > supported Compaq adapters utilize portions of the control protocol from > earlier" > > I'm specifically looking at the Proliant DL360, which has this card > ... can you provide any comments, or insight, concerning what the man > page states? Should I shy away from this controller? :(We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, and they seem to work fine with ciss driver. Just this damn bug (kern/83375) which is not related to ciss driver...> > Thanks ...-- Philippe Pegon
Hi, On Thu, Nov 24, 2005 at 07:01:17PM +0100, Philippe Pegon wrote:> We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, > and they seem to work fine with ciss driver. Just this damn bug > (kern/83375) which is not related to ciss driver...do you know of any method to monitor these controllers with FreeBSD, e.g. to detect drive failures? -- Sascha
Marc G. Fournier wrote:> > Hi .. > > Having read the man page, there is alot in there that makes me > wonder whether going with a HP Smart Array P600 is a wise idea ... > > "The Compaq ciss adapters require faked responses to get reasonable > behavior out of them. In addition, the ciss command set is by no > means adequate to support the functionality of a RAID controller, and > thus the supported Compaq adapters utilize portions of the control > protocol from earlier" > > I'm specifically looking at the Proliant DL360, which has this card > ... can you provide any comments, or insight, concerning what the man > page states? Should I shy away from this controller? :(The P600 should work just fine.
Sascha Holzleiter wrote:> do you know of any method to monitor these controllers with FreeBSD, > e.g. to detect drive failures? > >No, but the code is in the driver and can be easily adapted to a userland program to probe the controller through the ioctl interface, but the easiest way is to just have something monitor syslog.
On Thu, Nov 24, 2005 at 07:15:07PM +0100, Sascha Holzleiter wrote:> > On Thu, Nov 24, 2005 at 07:01:17PM +0100, Philippe Pegon wrote: > > We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, > > and they seem to work fine with ciss driver. Just this damn bug > > (kern/83375) which is not related to ciss driver... > > do you know of any method to monitor these [ciss] controllers with FreeBSD, > e.g. to detect drive failures?You could simply monitor the corresponding ciss syslog-messages and scan for state changes (ie from OK to something else). Apart from reboot-messages, you only get messages if states are changing... Regards, Holger Kipp alogis AG, Berlin
Holger Kipp wrote:> On Thu, Nov 24, 2005 at 07:15:07PM +0100, Sascha Holzleiter wrote: > >>On Thu, Nov 24, 2005 at 07:01:17PM +0100, Philippe Pegon wrote: >> >>>We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, >>>and they seem to work fine with ciss driver. Just this damn bug >>>(kern/83375) which is not related to ciss driver... >> >>do you know of any method to monitor these [ciss] controllers with FreeBSD, >>e.g. to detect drive failures? > > > You could simply monitor the corresponding ciss syslog-messages > and scan for state changes (ie from OK to something else). Apart > from reboot-messages, you only get messages if states are > changing...you could also use camcontrol(8) for example : camcontrol inquiry da0 -D> > Regards, > Holger Kipp > alogis AG, Berlin-- Philippe Pegon
Sascha Holzleiter wrote:> Hi, > > On Thu, Nov 24, 2005 at 07:01:17PM +0100, Philippe Pegon wrote: > >>We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, >>and they seem to work fine with ciss driver. Just this damn bug >>(kern/83375) which is not related to ciss driver... > > > do you know of any method to monitor these controllers with FreeBSD, > e.g. to detect drive failures?We use camcontrol(8) and a perl script. The perl script runs every five minutes a command like that : # camcontrol inquiry da0 -D pass0: <COMPAQ RAID 5 VOLUME OK> Fixed Direct Access SCSI-5 device and when the output changes, it sends an alarm by mail. -- Philippe Pegon
On Thu, 24 Nov 2005, Holger Kipp wrote:> On Thu, Nov 24, 2005 at 07:15:07PM +0100, Sascha Holzleiter wrote: > > > > On Thu, Nov 24, 2005 at 07:01:17PM +0100, Philippe Pegon wrote: > > > We have about thirty HP servers (DL360 and DL380) with FreeBSD 5 and 6, > > > and they seem to work fine with ciss driver. Just this damn bug > > > (kern/83375) which is not related to ciss driver... > > > > do you know of any method to monitor these [ciss] controllers with FreeBSD, > > e.g. to detect drive failures? > > You could simply monitor the corresponding ciss syslog-messages > and scan for state changes (ie from OK to something else). Apart > from reboot-messages, you only get messages if states are > changing...Maybe this is what you are looking for? http://people.freebsd.org/~jcagle/ /Bjorn