FS is frequently facing i/o errors

[Ramesh0408]

Hi All,

I am using storage foundation 6.1 with Solaris10 u10. FS are frequently facinf i/o errors. I am unounting the FS and mounting it back. It is mounting it back without any issues. but I would like to why it is happening regularly. Below output is showing when I am trying to get the FS version. Is it causing the issue? or why it is showing below unrecognized superblock. Please advice. Thanks in advance.

 fstyp -v /dev/vx/dsk/XXXX/XXXX|grep -i version
Unrecognized Superblock.  error = 5052
magic XXXX  version 7  ctime Wed Oct 03 20:35:59 2012

[Gaurav_S]

Are there any messages appearing in /var/adm/messages during same time ?  What storage it is ?

Is this an upgraded setup ? I see FS version 7 on SF 6.1 ?

G

[Ramesh0408]

Hi Gaurav,

Thanks for your reply. Please find below comments on your questions.

* Yes, I have messages. Please find below.

root@XXXX # cat /var/adm/messages|grep XXXXXXvol0

Jun 25 15:21:06 XXXXX vxfs: [ID 702911 kern.warning] WARNING: msgcnt 27 mesg 037: V-2-37: vx_metaioerr - vx_iasync_wait - /dev/vx/dsk/XXXXXX/XXXXXXvol0 file system meta data write error in dev/block 0/3856

Jun 25 15:21:06 XXXXX vxfs: [ID 702911 kern.warning] WARNING: msgcnt 28 mesg 031: V-2-31: vx_disable - /dev/vx/dsk/XXXXX_dg/XXXXXXvol0 file system disabled

* Storage is HITACHI.

* Yes, It is upgraded setup as 6.1 require Min. FS version 7 and DG version 160. We met both requirements. 

[Gaurav_S · Accepted solution]

Hi,

Can you post entire messages file for that time, metadata write error may be related to an underneath issue accessing the disk/storage

Do you have VCS in use or its a standalone system ? If VCS is in use, do you use IOFencing ?

Also, do mention exact model of Storage, would be able to check the recommended settings for Storage/host mode.

G

[Marianne]

I agree with Gaurav - we need to see all entries in messages file for that time frame. 
These type of issues are normally caused by hardware problems.

I suggest showing us everything between Jun 25 15:21:00 and 15:21:06.

We need to see if any errors are logged for the underlying hardware.