New backup system on 11.38 corrupt files

cPanelKenneth

cPanel Development
Staff member
Apr 7, 2006
4,607
79
458
cPanel Access Level
Root Administrator
Hi Kenneth,

Thank you for the update. I was given two case numbers in my ticket 67985 and 68049. Out of interest, what's the difference? Are they both responsible for the corrupt backups?
67985 and 68049 were duplicates of each other. We tend to merge newer cases into the older cases. 67985 is what will appear in the change log.

As for the comment about improved security, I speculate this thought stems from being able to disperse the backups across multiple systems. This reduces the likelihood of a malicious user deleting the backups.
 

Gareth-AWD

Well-Known Member
Jul 3, 2008
195
13
68
London, UK
cPanel Access Level
Root Administrator
This seems to have been resolved in 11.38.08. Yay

A big thanks to the Team and Darren in Development for getting this sorted.

Out of interest, why does the email for complete backups get sent BEFORE the FTP queue has finished uploading? I'm guessing the new backup system has a separate element that queues the files rather than sending them directly in the script.
 

cPanelKenneth

cPanel Development
Staff member
Apr 7, 2006
4,607
79
458
cPanel Access Level
Root Administrator
This seems to have been resolved in 11.38.08. Yay

A big thanks to the Team and Darren in Development for getting this sorted.

Out of interest, why does the email for complete backups get sent BEFORE the FTP queue has finished uploading? I'm guessing the new backup system has a separate element that queues the files rather than sending them directly in the script.
I am happy to hear your issue is resolved :)

Your guess is correct. The upload to the remote systems is handled by a separate process. the backup process generates the tarball(s), and stages them in a local directory. A secondary system copies from the local staging directory to all the configured remote systems.

Using a model like this allows generation of the backups to complete much sooner than with the older system. We should likely re-examine how the admin is notified of these events, so the admin has a clear understanding which stage of the backup is complete.
 

sp3ctre69

Well-Known Member
Aug 14, 2006
107
4
168
Hi, sorry to report that it's not entirely working for me. It's better than it was but I had a backup that was 2.5GB that only transferred half. It was to a slower FTP connection which could have had something to do with it, the 2nd FTP location got the full thing.
 

sp3ctre69

Well-Known Member
Aug 14, 2006
107
4
168
I'm happy that the faster of the FTP targets is working but still quite worried that a) the other one isn't, and b) It doesn't fail nicely, there is no sign other than digging in the logs that suggests it failed. I have used slow FTP targets on the legacy system for a long time and never had any timeout issues.
 

sp3ctre69

Well-Known Member
Aug 14, 2006
107
4
168
What does this post mean? Are you saying that you think your system downgraded over night?
Hi, sorry it wasn't clear.

When I checked the changelog (looking for 11.38.9) this morning the latest listed was 11.38.7. Yesterday there was a 11.38.8 (which is on my server). I just wondered why it was pulled from the changelog.

Confusingly I just checked back again and 11.38.8 is now back on the changelog.

Jim
 

sp3ctre69

Well-Known Member
Aug 14, 2006
107
4
168
Cool, thanks.

Do you have any info on what the issue with the backups is? I am being told on my support call it was something to do with "out of memory" but I understood it to be a timeout issue? Either way 11.38.8 didn't fix it so I could do with knowing if it is still being looked at.

Thanks
 

Infopro

Well-Known Member
May 20, 2003
17,090
518
613
Pennsylvania
cPanel Access Level
Root Administrator
Twitter
Thanks for that. We should leave that ticket to Technical Support I think.

In the meantime, I note you stating that local backups do work, and, you don't have any valid backups right now with this remote issue. If you haven't yet, I'd switch back to local if possible, for obvious reasons.

If it was me, I'd rename cpbackup directory (local) and generate fresh backups (if you can afford the space). Just to be sure I have up to date backups without issues. Once I was sure I had those, kill that renamed cpbackup directory.
 

cPanelMichael

Administrator
Staff member
Apr 11, 2011
47,908
2,216
463
Hi Everyone
I have the same problem as above - any solution??
I recommend opening a support ticket if you continue to experience the original issue reported on this thread. Please let us know the ticket number and we can update this thread with the outcome.

Thank you.