Backup freezes and hangs for days

Hi, I've got 64bit version of Centos 6 on VPS Virtualmin has scheduled backups daily. What happens is these processes hung forever without any outcome. And when they reach 3 (which is limite set for this module - I get error "There are currently 3 backups running, and the system-wide limit is 3") What is the problem? - Some hang there for 3-5 days and I have to kill them manually. It does not backup due to this problem. Is it something to do with VPS itself? There is plenty of memory resources etc, so should not be a problem. Even when i do it manually - it does the same thing - in the browser output it goes to certain point where it does not respond.

Status: 
Active

Comments

Howdy -- are any of those backups currently running on your server (that is, are the processes still there)? If so, could you attach the output of this command:

ps auxwf

Also, what is the destination of those backups? Is it a local directory, or is it sending them to a remote FTP/SSH/S3 server?

yes, still running
also backup is going to the external ftp, one file per domain (tar.gz) which is also giving strange log:
backs up 2 domains, on third one (only 22Mbytes size) it quits
"upload finished. user disconnected gracefully."
here is the ps auxwf:
see attachment

If you try a backup to the same destination from the command line with the virtualmin backup-domain command, does it complete successfully? And if not, how far does it get?

Hmm, there don't seem to be any backup related process running at the moment... if you go into Backup and Restore -> Backup Logs, what do the backup logs look like for a backup that hung up?

this one here

root 28921 0.0 2.5 190208 107220 ? S 14:47 0:02 \_ /usr/libexec/webmin/virtual-server/backup.cgi

and in log it says the same
There are currently 3 backups running, and the system-wide limit is 3
Deleting backups from /backup_domains/%d-%m-%Y on FTP server xx.xxx.xxx older than 14 days ..
.. no backups to delete were found

Final status Failed

will try backup in console and will let know asap

backs up locally very quickly and no errors very strange.. then it must be something to do with the ftp then...? does it have some sort of time outs set - the scheduled virtualmin script when it does the backup? It looks like most likely cause is the ftp right now..

My apologies, I missed the backup process in the "ps" output above -- I do indeed see it now, but I unfortunately don't see anything that helps diagnose the issue in the processes under it.

The FTP server you're using -- does that server also run SSH?

If so, would it be possible to use that?

Sometimes, FTP can be "funny" with the way it handles ports and such, where SSH is a bit more straight forward. If you have the option, I'd be curious if the backups work properly when using an SSH account as the backup destination.