Operating System - Tru64 Unix
1748092 Members
5901 Online
108758 Solutions
New Discussion юеВ

Re: AdvFS I/O Error

 
Srivathsan
Frequent Advisor

AdvFS I/O Error

Hi,

I am running an Alpha Server 4100 with Tru64 V4.0D ( Sorry, Can't upgrade as of now !! )

The server has been crashing on and off every weekend.

The error reported in binary.errlog say the following

********************************* ENTRY 127. *********************************
----- EVENT INFORMATION -----

EVENT CLASS ERROR EVENT
OS EVENT TYPE 302. PANIC
SEQUENCE NUMBER 156.
OPERATING SYSTEM DEC OSF/1
OCCURRED/LOGGED ON Sat May 15 22:35:09 2004
OCCURRED ON SYSTEM arolia2
SYSTEM ID x00070016
SYSTYPE x00000000
PROCESSOR COUNT 4.
PROCESSOR WHO LOGGED x00000003
MESSAGE panic (cpu 3): del_xtnt_array: Can't
_ref BMT page for chain mcell

********************************* ENTRY 128. *********************************
----- EVENT INFORMATION -----

EVENT CLASS ERROR EVENT
OS EVENT TYPE 199. CAM SCSI
SEQUENCE NUMBER 153.
OPERATING SYSTEM DEC OSF/1
OCCURRED/LOGGED ON Sat May 15 22:35:09 2004
OCCURRED ON SYSTEM arolia2
SYSTEM ID x00070016
SYSTYPE x00000000
PROCESSOR COUNT 4.
PROCESSOR WHO LOGGED x00000003

----- UNIT INFORMATION -----

CLASS x0000 DISK
SUBSYSTEM x0000 DISK
BUS # x0003
x00CB LUN x3
TARGET x1

And the machine's /var/adm/messages file reports AdvFS I/O Errors



May 15 22:33:34 arolia2 vmunix: AdvFS I/O error:
May 15 22:33:34 arolia2 vmunix: Domain#Fileset: r2_domain#appl
May 15 22:33:34 arolia2 vmunix: Mounted on: /appl
May 15 22:33:34 arolia2 vmunix: Volume: /dev/rzd25c
May 15 22:33:34 arolia2 vmunix: Tag: 0x000016cd.8001
May 15 22:33:34 arolia2 vmunix: Page: 3
May 15 22:33:34 arolia2 vmunix: Block: 2226096
May 15 22:33:34 arolia2 vmunix: Block count: 16
May 15 22:33:34 arolia2 vmunix: Type of operation: Write
May 15 22:33:34 arolia2 vmunix: Error: 5
May 15 22:33:34 arolia2 vmunix: To obtain the name of the file on which
May 15 22:33:34 arolia2 vmunix: the error occurred, type the command:
May 15 22:33:34 arolia2 vmunix: /sbin/advfs/tag2name /appl/.tags/5837

The machine rebooted and filesets from 2 domains didn't mount.

Had to doa manual reset of the HSG70 Controllers to make this work.

Any help please !!
5 REPLIES 5
Orrin
Valued Contributor

Re: AdvFS I/O Error

Hi,

Looks like you have a hardware problem. Look at disk - bus 3 lun 3 target 1.

Regards,
Orrin.
Mohamed  K Ahmed
Trusted Contributor

Re: AdvFS I/O Error

You have a problem communicating with the disk in question, it maybe a bad hard disk or a problem connecting to it. The OS puts it as an error and you have to clear the error before you can mount it back.
There are some commands relating to the AdvFS that will fix that for you (if there is no hardware problem), you can find the commands in /sbin/advfs.

Indepth analysis would depend on your configuration and your disknames

Mohamed
Michael Schulte zur Sur
Honored Contributor

Re: AdvFS I/O Error

Hi,

you could do a test for bad disk
dd if=/dev/rrd25c of=/dev/null bs=1024k
umount domain in question and check with
/sbin/advfs/verify domain
I don't think, resetting the controller will do you much good, if the fileset is damaged. If the disk in question mirrored?

greetings,

Michael
Yong_7
Frequent Advisor

Re: AdvFS I/O Error

Hi,

this is telling me that machine has the problem to write to disk (3/3/1) through HBA.

the message in /var/adm/messages you put here indicates the write error, is there any domain panic message around ? take a look at SysAdmin pdf file for more details.

when it crashs again, boot to single user,
#/sbin/advfs/verify -f r2_domain

if the domain can't be mounted,
#/sbin/advfs/salvage

is your friends.

after that, if i were you, i would recreate the domain on another disk and vdump/vrestro
to relocate the data if data is critical.

Cheers !

YJ
Ravi_8
Honored Contributor

Re: AdvFS I/O Error

Hi,

seems to be problem in disk. hence the advfs i/o error, replace the disk which solve your problem
never give up