Vgcfgbackup/vgextend err

[Roger Baptiste]

Hi,

I got the following errors while doing a series of vgextends through a script on a VG:
(thats because the vg has 40 PV's and i was running vgextends with four pvs in a loop)
***
Volume Group configuration for /dev/tempvg has been saved in /etc/lvmconf/tempvg.conf
Volume group "tempvg" has been successfully extended.
vgcfgbackup: IO error on configuration file: Too many open files
Volume group "tempvg" has been successfully extended.
vgcfgbackup: Cannot open the control file "/dev/tempvg/group":
Too many open files
*********

The Vg did get created/extended correctly and i can see that on the vgdisplay. I ran vgcfgbackup manually again and it went through fine.
Just curious to know, what this error means?

The VG is 540 Gb size. (temporary vg).

thanks
-raj

[Scott Marer]

It would appear that the max number of open files was reached when trying to extend the volume group.

I suggest that you look in the syslog.log to see if any adidtional errors exist.

[A. Clay Stephenson]

Without knowing what your script was doing, it's difficult to say for sure. That error message is associated with EMFILE - not enough file descriptors rather than ENFILE - system wide file table exhausted. I would check maxfiles and maxfiles_lim.

[S.K. Chan]

The maxfile/maxfile_lim got hit by the limit ?

[Roger Baptiste]

The script is nothing but a series of vgextends.

eg: vgextend tempvg pvname1 pvname2 ...
vgextend tempvg pvname5 pvname6 ..
and so on for the same tempvg.

The vg seems to have extended fine , since the vgdisplay -v shows all the pv's. My doubt is about the error message?

I think, after every vgextend, a vgcfgbackup is automatically done and the error occured during this time; because of successive vgextends, there could have been a overlap of one vgcfgbackup running while a new one started.

I am not sure, what this has to do with Kernel parameters, because this is a simple sequence of 10 vgextends on the same VG, on an empty system.

thanks
-raj

[pap]

Hi Rajman,

Just do following things and I am sure you will resolv your problem.

1. check your maxvgs parameter in kernel (use kmtune | grep maxvg ) if it is lessa than 40 it will not allow you to create more group files for more expected vgs. So just increase that parameter to 100 and rebuild the kernel. You will definately solve your problem.

If at all you are unable to solve your problem, check the unique minor number among along all group files of all vgs .

Good luck.

-pap