"archive" for backups
Christopher J. Miller
via
How should the "archive" program be used to backup the SMB database... should I run at as the saffron user or root? Seems to fail when I run it as saffron, but it completes properly as root. Also, should it be run with the "-b" switch to backup all nodes in the cluster?
Comments are currently closed for this discussion. You can start a new one.
Support Staff 2 Posted by David E. Young on 04 Aug, 2009 02:38 PM
I'm able to run archive as a normal saffron user. What exactly do you mean "seems to fail"?
Yes, to back up all nodes use the -b flag.
3 Posted by Christopher J. Miller on 04 Aug, 2009 02:45 PM
It outputs a lot of messages showing some commands returned a non-zero exit status. I'm not sure of this means that some portion of the process is failing, but I checked the archive size and it does seem to match previous ones that were run as user root. So it would appear that maybe it is completing, but not without some problems.
Here is some of the output, there is a lot more of this, I just clipped the last part of it to paste here.
Command 'ssh 1363D1App1 chmod go+w /tmp/mysql/saffron_at_simulation_community_1 ' returned non-zero exit status 1
Command 'ssh 1363D1App3 chmod go+w /tmp/mysql/saffron_at_simulation_community_3 ' returned non-zero exit status 1
Command 'ssh 1363D1App2 chmod go+w /tmp/mysql/saffron_at_simulation_community_2 ' returned non-zero exit status 1
Command 'ssh 1363D1DB2 chmod go+w /tmp/mysql/saffron_at_simulation_community_0 ' returned non-zero exit status 1
Command 'ssh 1363D1App2 chmod go+w /tmp/mysql/saffron_at_simulation_community_2 ' returned non-zero exit status 1
Command 'ssh 1363D1DB2 chmod go+w /tmp/mysql/saffron_at_simulation_community_0 ' returned non-zero exit status 1
Command 'ssh 1363D1App3 chmod go+w /tmp/mysql/saffron_at_simulation_community_3 ' returned non-zero exit status 1
Command 'ssh 1363D1App1 chmod go+w /tmp/mysql/saffron_at_simulation_community_1 ' returned non-zero exit status 1
Support Staff 4 Posted by David E. Young on 04 Aug, 2009 02:52 PM
So chmod is complaining about setting permissions on the /tmp/mysql/* files. Look in the node-specific logs (logs//messages) to see exactly why chmod is unhappy.
I do NOT see those complaints when I run archive on our clusters here, so this must be environmental...
5 Posted by Christopher J. Miller on 06 Aug, 2009 04:05 PM
I believe I figured out what the cause was. The first time "archive -b " was run, I ran it as root. So the /tmp/mysql directory on each node was created with root:root ownership. Once I chowned it back to saffron:saffron, it appears the archive was able to complete when run as userid "saffron".
Support Staff 6 Posted by David E. Young on 12 Aug, 2009 12:45 PM
Yup, that would sure do it. Good job figuring it out.
David E. Young closed this discussion on 12 Aug, 2009 12:45 PM.