02-21-2014 07:09 AM
Urgent...!!!!
While runnung script the veritas command vxdiskadm -o allgs got stuck. My OS is running VVM and VCS.
I tried to run manually and it is hanging. /var/adm/messages showing below error messages.
My peers suggetion is to reboot the server. Can you suggest any other way to resolve this issue?
vxdmp: [ID 443116 kern.notice] NOTICE: VxVM vxdmp V-5-0-0 i/o error occured (errno=0x0) on dmpnode 267/0x2842
02-24-2014 04:37 AM
Probably vxdisk and not vxdiskadm?
I would expect additional device errors in /var/adm/messages if there are hardware and/or I/O issues.
Please copy messages file to messages.txt and upload as File attachment.
02-24-2014 05:39 AM
I would say, look at the current vx processess running, also how is server performance ?
It is quite possible that some of vx command is spawning & has gone into a loop or too many vx process are running. Any other vx tasks syncing in background ?
# ps -ef | grep -i vx
# vxtask list
# sar 5 5
# prstat -l
G
02-25-2014 07:32 AM
Hi Marianne,
Thanks for reply, That is my typo it vxdisk only. I can't upload messages file bcz of security issues but I can say there are lot of transport errors on LUNs. find below output from iostat -en. I have lot of luns like below with transport errors.
0 0 16212 16212 c5t60060E80164D5130d99
0 0 16213 16213 c5t60060E80164D5130d98
0 0 0 0 c5t60060E80164D5130d89
0 0 16213 16213 c5t60060E80164D5130d47
02-26-2014 03:57 AM
... but I can say there are lot of transport errors on LUNs.
Without seeing exact details of the errors, it sounds like a hardware issue.
Are errors only on disks?
Or on hba's as well?
You need to carefully look at errors in messages file and determine the source of the issues.
Check SAN/switch monitoring tools as well.
Depending on where errors are originating from, you will need to log a call with the relevant hardware vendor.
At this point in time DMP is reporting the issue, not causing it.
02-26-2014 03:58 AM
Raised call with Symatec. Even they suugeted to reboot.
Thanks for your replies Marianne and Gaurav.
02-27-2014 07:21 PM
I think call to hardware vendor(s) would be more appropriate...
DMP was merely reporting the issue, not causing it.