Error: The AdvFS domain is inaccessible (paniced)

[nasimuddin ansari]

Dear All
We have an ES40 running Tru64 ver 5.1A with latest patch kit. This server a quad CPU server having 2 GB memory. It is attached with direct storage RA3000 on PCI based SCSI card QLOGIC ISP1040B/V2 - Differential Mode. it has been configured for 02 virtual disk in RAID 5 .virtual discs has two domains domain_dsk1b & domain_dsk2b. We are getting below error for domain_2b having only one fileset named netscape.


AdvFS I/O error:
A read failure occurred - the AdvFS domain is inaccessible (paniced)
Domain#Fileset: domain_dsk2b#netscape
Mounted on: /usr/netscape
Volume: /dev/disk/dsk2b
Tag: 0x00013f56.80cf
Page: 0
Block: 7109872
Block count: 16
Type of operation: Read
Error: 5 (see /usr/include/errno.h)
EEI: 0x300
AdvFS initiated retries: 0
Seconds from first I/O attempt to this failure: 0
Total AdvFS retries on this volume: 0
To obtain the name of the file on which
the error occurred, type the command:
/sbin/advfs/tag2name /usr/netscape/.tags/81750

When we received this message, we are not able to access the domain. After reboot, domain is again accessible & usable. Within one or two days, same problem repeated again.

In attached file output of "vmstat 1 5" and "sysconfigdb -l" are given for analysis. Any help & suggestion appreciated. TIA

[Venkatesh BL]

Please post the output of "/sbin/advfs/tag2name /usr/netscape/.tags/81750"

[Michael Schulte zur Sur]

Hi,

please check disk for read errors with
dd if=/dev/rdisk/dsk2b of=/dev/null bs=1024k

greetings,

Michael

[Joris Denayer]

Hi,

The error is returning after a couple of days.
Then the errorlog must be investigated for HW errors on the devices. In this case -> dsk2.


If no hardware events are logged just before the Domain Panics, you might have a logical corruption. Run /sbin/advfs/fixfdmn on the Domain that just paniced.
In case of irrecoverable corruptions, you should recreate the domain#fileset and restore a backup.


Good Luck
Hope this helps

[Ivan Ferreira]

You should run the /sbin/advfs/verify command on the domain. If the domain is inaccessible, you can try with fixfdmn. As last option, you can use salvage to recover the data.

If the binary errlog does not have disks events at the time of the failure, check with

dia -R -i disk

[nasimuddin ansari]

Thank to all of you for quick reply.

I have recreated the domain and fileset on the harddisk. Since then I am not getting error.

If problem persisit, I will do other things as suggested in post.
TIA