topic Re: vxfs errors in Operating System - HP-UX

vxfs errors

brian_31 — Thu, 26 Oct 2006 15:25:11 GMT

I have the following vxfs errors in syslog
msgcnt 3 vxfs: mesg 001: vx_nospace - /dev/data/sap1 file system full (1 block extent)
msgcnt 5 vxfs: mesg 038: vx_dataioerr - /dev/data/sap1 file system file data write error
msgcnt 6 vxfs: mesg 033: vx_check_badblock - /dev/data/sap1 file system had an I/O error, setting VX_FULLFSCK

What does this mean? would i have worry given that the message came right after the filesystem was full?

Thanks

brian

Re: vxfs errors

Yang Qin_1 — Thu, 26 Oct 2006 15:29:54 GMT

It say your file system mounted with /dev/data/sap1 is full. Your applications which using that file system failed because that.

Run:

bdf /dev/data/sap1

to identify which file system is full and clean it ASAP.

Yang

Re: vxfs errors

brian_31 — Thu, 26 Oct 2006 15:45:35 GMT

as i said this filesystem was full and is now ok. does this error post any worries? what does setting vx_fullfsck mean?

Thanks

brian

Re: vxfs errors

Patrick Wallek — Thu, 26 Oct 2006 15:58:29 GMT

VX_FULLFSCK means that next time fsck is run on this LV, a FULL fsck, 'fsck -o full', will be run.

I am honestly not sure about the vs_dataioerr data write error. It may be that since the filesystem filled up the app couldn't write any more, but I have never seen a full filesystem generate that particular error.

If you have a support contract for this machine I would be tempted to open a call and quiz HP on this.

Re: vxfs errors

Yang Qin_1 — Thu, 26 Oct 2006 16:12:46 GMT

You had:
1. mesg 001 -- file system full and it cause
2. mesg 038 -- file data write i/o error and because this i/o error you got
3. mesg 033 -- the kernel set the VX_FULLFSCK flag as a precautionary measure.

Now you have cleaned that file system. If you believe that all applications which access that file system is running OK. You can leave them running and keep a close eye on them. If you find that you have problem to access some files / directories. Then you have to "umount /dev/data/sap1" and run "fsck" to check / repair that lvol.

It may help if you run "lvdisplay -v /dev/data/sap1" To see if there is some errors in that lvol. You can also check disks/LUNs used by that lvol by running "pvdisplay -v /dev/dsk/c#t#d#"

Yang