Re: Need help identifying error: A_RSCheckCond interrupt on ha=0 Steve M. Fabac, Jr. wrote:
> I received the following e-mail from a client:
>
> > last two nights I have had a WARNING on the server
> >
> > WARNING: A_RSCheckCond interrupt on ha=0 id=6 lun=0 tag=00
> > CPU2: WARNING: A_RSCheckCond interrupt on ha=0 id=6 lun=0 tag=00
> >
> >
> > repeated multiple times
> >
> > It looks like it's occurring between 1:15 A.M. and 3:30 A.M. when the backup is running.
> >
> > It completed backup last night, but didn't the previous night.
>
> Searching google comes up with three foreign language hits,
> one implicates oss478a, one LSI and SCO 5.0.x.
>
> Searching SCO's knowledge base for A_RCheckCond turns up zip.
> Searching Adaptec.com and LSI turns up zip.
>
> The client is running a SuperMicro X5DL8-GG Dual Xeon
> box with SCO 5.0.6 and two CPUs. The hard disk is a
> RAID-5 on Adaptec 2010 ZCR controller and the backup tape
> is an Ultrium tape used to replace a IOmega SCSI REV
> drive when the REV drive failed (would not eject media).
> Backup software is BackupEdge 2.2.
>
> The Ultrium tape is on an LSI Logic 53C895A SCSI controller.
>
> The system was installed on 3/17/2005 and has not produced
> this error in the past.
>
> Any ideas?
# cd /etc/conf/pack.d
# find * -name '*.[chaos]' | xargs strings -Fa | grep A_RSCheckCond
slha/Driver.o: A_RSCheckCond interrupt on ha=%d id=%d lun=%d tag=%b
slha/Driver.o: A_RSCheckCond_Used
"slha" is the driver for LSI Logic fka Symbios Logic nee NCR
Microelectronics 53c8xx SCSI HBAs.
I google'd the string and found 4 hits, none of which is a "foreign
language hit". They are all strings of gibberish extracted from the
middle of binary archives; all are archives which happen to contain one
version or another of the OSR5 "slha" driver.
So, something is going wrong with either the HBA or the tape drive. The
HBA is much easier to replace (no investment in tapes & data), I
recommend swapping in a cheap supported HBA which can talk to whatever
SCSI bus the drive has.
If you go to a different driver, there's a chance that even if it's a
drive problem, you will get better diagnostics out of the driver. The
string you're getting includes "CheckCond", short for "Check Condition",
a SCSI protocol message (see wikipedia.org/wiki/SCSI_check_condition).
One might hope that the driver would follow up by asking "What
condition?" and reporting that to you.
Swap in an HBA and be prepared to potentially need to swap in a
compatible drive.
If you have any sort of diagnostic software for the Ultrium drive, run
it. If you have cleaning tapes, clean it.
>Bela< |