SUSE 10 on Proliant DL580 cciss: cmd has CHECK CONDITION

[davep75]

I have a Proliant DL580 G5 running SUSE 10 Linux. Occassionally I see an error in /var/log/messages (and dmesg):

cciss: cmd ffff810008180000 has CHECK CONDITION byte 2 = 0x3

The host does not crash but it is having a problem with intermittant hanging. (Although not when this error appears).

The root disk (/dev/cciss/c0d0) is actually a pair of internal disks mirrored by SmartArray.

Any help would be appreciated. Thanks!!

[Ciro Iriarte]

If you installed the Prolian Support Pack, try:

hpacucli ctrl slot=0 show status
hpacucli ctrl slot=0 ld all show status
hpacucli ctrl slot=0 pd all show status

[davep75]

The hpacucli commands show everything ok. I occassionally see the following in /var/adm/messages:

Jun 6 14:51:37 gporac3 kernel: hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }
Jun 6 14:51:37 gporac3 kernel: ide: failed opcode was: unknown
Jun 6 14:51:37 gporac3 kernel: hda: drive not ready for command
Jun 6 14:52:39 gporac3 kernel: ide-cd: cmd 0x3 timed out
Jun 6 14:52:39 gporac3 kernel: hda: irq timeout: status=0xd0 { Busy }
Jun 6 14:52:39 gporac3 kernel: ide: failed opcode was: unknown
Jun 6 14:52:39 gporac3 kernel: hda: ATAPI reset complete

[Ciro Iriarte]

That seems to be a optical unit (cd/dvdrom). SLES10SP2 is out, is an upgrade a option?

[davep75]

I may try an upgrade to fix the intermittant hanging problem, if the new version is supported by the application vendor (Oracle).

[davep75]

The problem turned out to be the missing 2nd set of drives. We had two sets of 2 mirrored drives, each with a different OS/app stack. We either had one, or the other, set plugged in at any given time; except when the first OS image was built. So the controller was concerned about the 2nd set of drives being offine (even though we didn't need them). Not sure if that makes sense but the bottom line is that the error was annoying, but harmless.

[Maaz]

Hi
>The problem turned out to be the missing 2nd set of drives

in /etc/fstab add the "hotplug" in the option

e.g
/dev/sdb /mp fs options,options,hotplug 0 0

Regards

[SERKAN AKÇIN]

Hi davep75,

Yesterday we get the same kind of message on Proliant DL585 G1 with RHEL5.2_x86_64.
Message like...
Jan 8 00:22:48 gordion kernel: cciss: cmd ffff8100f5700000 has CHECK CONDITION byte 2 = 0x3
Jan 8 00:22:48 gordion kernel: cciss: cmd ffff8100f570a350 has CHECK CONDITION byte 2 = 0x3
J

We are searching the problem. But nothing found yet.

After we reboot the server, at the smart array controller boot process, something writen like: data is valid, and recovering...

Now all the system is stable.

But we don't know the start of the problem.

Did you find a solution of your problem.