I’ve seen other users here have had this problem as well (judging from the archives) but there has never been a solution posted. I believe that I have found the answer to this problem, somewhat accidentally:
In the process of setting up a script that would periodically check the BackupPC log for the ‘BackupPC::Lib->new failed’ error, I needed to do several restarts of the server. Apache wasn’t starting on its own giving ‘nodename nor servname provided, or not known: mod_unique_id' in the error log. Simply commenting out this module solved the apache startup problem … and also, apparently, the ‘BackupPC::Lib->new failed’ error as well. The script I wrote has not once fired a restart procedure since I commented out the module. Previous behavior was that the error would crop up at least twice a day and persist until I restarted BackupPC.
Now, I’m not quite certain what mod_unique_id does but I haven’t been using it in any other apache installation so I haven’t been too worried to do the research. If anyone would kindly explain any portion of this problem, I’d be much appreciative. I plan on writing a howto for BackupPC for Debian and Mac OS X and this piece of info would be pretty handy.
B
I’ve recently needed to reinstall BackupPC on Mac OS X 10.5 Leopard. While it had worked previously on this platform, I’m getting some very odd errors that driving me insane:
I’m able to run backups for a period of time but some point in time during the day I’ll begin to get repeated iterations of the following:
/opt/local/BackupPC/bin/BackupPC_dump: Wrong user: my userid is 501, instead of (tlctoor)
Please su tlctoor first
BackupPC::Lib->new failed
After I get this, newly initiated processes will fail but previously active processes seem to continue. I then need to restart the BackupPC service to get backups back on track.
What am I missing, where should I look?
Brian
