These forums are locked and archived, but all topics have been migrated to the new forum. You can search for this topic on the new forum: Search for SSH scheduled backup not working but manual run is?! on the new forum.
I have a SSH server running on a different port to 22 - this is the server I back up to
using server:port
When I invoke the backup manually - it works fine
But when the schedule runs it fails with the error
Failed to connect to SSH server : ssh: connect to host SERVER port xxxxx: Connection timed out
lost connection
Any ideas why this might work for one and not the other??
Howdy,
Hmm, I'm not sure. Is there any chance you could post a screenshot of what your Backup and Restore -> Scheduled Backup -> BACKUP screen looks like?
I'd like to see the various settings you have setup in there.
-Eric
Sure see attachment
btw im using ssh key exchange for password-less login - hence no user/password
Sorry missed these bits
Okay, and just to verify -- when you run the backup manually, how are you going about that?
-Eric
I run it from the Scheduled backups list main screen - I click backup next to the job - then I click backup now or backup in the background
Bump.
Is it possible to invoke the manual run from cron?
One last question, and we may need to get Jamie involved to look into it further...
But when you receive the error message when the backup runs -- it appears to mention a server name and port number in that error message.
Is the port number it mentions correct?
Or is it reverting back to trying to use port 22?
-Eric
Yep the server and port are correct in the message. I also turned on the email notification and its the same in there. Whether virtualmin is actually using that is another question.
Like I said what's weird is the same job (same settings) when run manually, works. Just not when scheduled.
It might be worth trying running
tcpdump
when the backup is running to see what port it is actually using. This may be a Virtualmin bug (using the wrong port), or it may be that the network / firewall is setup differently at the time the scheduled backup runs.''
Wow do I feel stupid.
I did what you said Jamie - TCPDUMP set a schedule for timenow + 1 min watched tcpdump like a hawk - and the backup went through!
I was like WTF - then realised a few months back I'd had a lot of SSH attacks - so to counteract this I had setup an allow access schedule(the backup was outside that schedule to not affect my home bandwidth) for any IP - but didn't exclude my webserver from the range of IPs
Just to confirm it was my fault and no fault of virtualmin!
Sorry Rob