Backups deadlock

Hi,

Could you send a copy of the backup log?
This will be too long to post here, but you can upload it to an online service (such as https://pastebin.com/) and post a link to the content here

Best Wishes,
David

This is the first process: link.

Then, I have 4 more logs like this: link.

Hi Ivan,

Is that the entire log? It looks like various bits have been cut out of it.

David

No, I hide my domain and server IP. Also, since the list of tables was big, I only left the first and the last one.

The first process ends at:

0123.619 (0) Table wp_wfReverseCache: Total rows added: 5 in 0.12 seconds

I set in my settings daily DB backups and weekly file backups. Last night, it should have done both but database backup failed again.

The “uploads” and “others” zip backups were successfully made and upload to Google Drive. I don’t know what it’s happening with DB backups.

Information from the logs:

• First process started at 04:29:39
• Deleted old temporary files (from the unfinished backup of the day before)
• The last entry in the log was:
  Table wp_revslider_sliders: Total expected rows (approximate): 0
• New process at 04:30:38
• Failed to gain semaphore lock
• New process at 04:31:27
• Failed to gain semaphore lock
• New process at 04:56:28
• Started files backup
• Stopped. Last log: “500 files added”
• Restarted at 05:08:38
• Stopped. Last log: “800 files added”
• Restarted at 05:11:35
• Restarted at 05:13:17
• Finished zip file
• Restarted at 05:19:29
• Stopped. Last log: “Googledrive chunked upload: 3.5 % uploaded”
• Restarted at 05:27:58
• Stopped. Last log: “Googledrive chunked upload: 56 % uploaded”
• Restarted at 05:29:53
• Stopped. Last log: “Googledrive chunked upload: 57.1 % uploaded”
• Restarted at 05:42:27
• 2883.310 (4) The backup apparently succeeded and is now complete

Hi,

When the database backup appears to stop, I’d suggest checking your PHP error log, to see if anything gets logged there at the time.

David

I checked that but I forgot to mention it. The error_log files had no errors related to this or at the time that backups failed.

Currently, I have db backups for the last two days: October 25 and 26. However, I would like to find the root of this and make sure that backups will always be made.

I had some unexpected behaviors with other plugins caused by W3TC object and database cache. Could that affect your db backup implementation?

Hi Ivan,

If your log file finishes with lines that begin with (0), then that indicates that your WordPress scheduler is not functional. After (0) there should be (1), then (2), up to at least (9) (as long as you leave it long enough). Please see:
https://updraftplus.com/faqs/my-scheduled-backups-do-nothing-backup-now-stops-midway/

It’s normal for the webserver to kill off a process after some time. But, it should be resumed by the WP scheduler later, which doesn’t seem to be happening.

However, in the absence of a complete log file, I’m not certain about this, as I don’t have the complete information in your case. If you can link the complete log file, that would help.

David

If you give an email, I’ll send you the log file. However, as I already posted, the last line of the log is:

0123.619 (0) Table wp_wfReverseCache: Total rows added: 5 in 0.12 seconds

You can send the log to dnutbourne(at)updraftplus.com.

Running through the FAQ above will at least rule out a number of possibilities, so we would suggest following that guide in the meantime:

https://updraftplus.com/faqs/my-scheduled-backups-do-nothing-backup-now-stops-midway/

I’ve just sent you the log files.

From the FAQ, the only thing I’m not sure is “loopback connections”. However, the cron jobs are called by WordPress but they fail and throw the message:

(0) Failed to gain semaphore lock (d) - another backup of this type is apparently already active - aborting (if this is wrong - i.e. if the other backup crashed without removing the lock, then another can be started after 3 minutes)

What is more, some days it works fine. So, I don’t think that could be the reason.

Hi,

The logs ending with the ‘Failed to gain semaphore lock’ are from backups that were fired at the same time as an already running backup, so they are terminated.

Could you try installing the WP Crontrol plugin? Then go to Tools->Crontrol and sending me a screenshot of that page? This will confirm if there is a problem with your scheduler.