savevg

[desbo]

We have just started to use the savevg command excluding all files to provide a way of recovering our volume group structure during a system reload. The file created by savevg is OK and a restore using restvg has been tested.

But when we run the savevg command we get the foloowing error
Creating information file for volume group mtmsvg......................./usr/sbin/lsjfs[48]: 95666 Killed
/usr/sbin/lsjfs[48]: 95670 Killed
/usr/sbin/lsjfs[48]: 26442 Killed
/usr/sbin/lsjfs[48]: 351718 Killed

we get 35 of these Killed messages.

If I run lsjfs command manually it works OK.

Why should this happen?
Is it important?

 

[donttreadonme]

The message "Creating information file for volume group..." is coming from /usr/bin/mkszfile which is called from mksysb which savevg is a link to.

The is a loop in the mkszfile script that gets the LV name, say hd2, and then gets the file system name in a for loop. This data is printed out to the image.data file. Farther down in the script the lsjfs is called again to get the aggregate size which is the 17th field in the lsjfs output.

If you look at the image.data file it shows at the end the commands for fs_data are /usr/bin/df and /usr/sbin/lsjfs.

Maybe? it is a problem with creating the image.data file. I am guessing that it prints out 35 times because you have 35 filesystems/logical volumes in the volume group.

 

[desbo]

yes, the number of Killed matches.

I have done some more testing and found that if I execute the command from the command prompt or by using "at now" the Killed messages do not appear.

However if i run it as a cron the Killed messages appear.

I have done it several times and it is consistant.

Why should running as a cron job be any different?

Under all circumstances I am running job as root user.