Annoying SCSI-error in syslog

[Jonas Petersson]

Hello

We have a HP-UX 11.i server running MC/Serviceguard 11.14 who had two local Auto-raids connected (VA 7100) We have gotten rid of those auto-raids and connected the cluster to an EVA5000-SAN. However, the syslog is still reporting the following over and over again

As far as I can see the SCSI errors reffers to the devices c8t0d1 and c4t0d1 which are two of the old VA 7100 disks.

See attached file for syslog entry.

Why is it still logging error messages from disks that are not connected to the system anymore?

[Joseph Loo]

hi,

since the disk are no longer there, do you mind removing those device files, i.e /dev/rdsk/?

also, please confirm it first before removing the device files.

# pvdisplay /dev/dsk/c#t#d#
# strings /etc/lvmtab

hope it help.

regards.

[Johan Lorimier]

Hi,

have you removed all references to the old disk ?
You have to remove it from the VG structure and from the ioscan output because cluster daemon scan all scsi path.

Johan

[Cheryl Griffin]

For each of the devices listed in the message "dev: bc080100", determine if the device exists on your system:
# ll /dev/* |grep 080100
# ioscan

If the device is bogus, use rmsf to remove the devices.

[Jonas Petersson]

Thanks for the fast respons in this matter. The answer to the problem was rmsf -a /dev/(r)dsk/cxtxdx. All the devices referensed by the SCSI error were removed when we moved from the VA 7100 to the EVA5000, however the device files hadn't been removed

[Joseph Loo]

hi jonas,

so do we get any points for helping?

it seems your track record for assigning points is not very good, i.e. 0 out of 7 assigned.

regards.