Support

Please look at the bottom of this page (under Support Policy Summary) for our support policy summary, containing important information regarding our working hours and our support policy. Thank you!

EXTREMELY IMPORTANT: Please attach a ZIP file containing your Akeeba Backup log file in order for us to help you with any backup or restoration issue. If the file is over 2Mb, please upload it on your server and post a link to it.

Description of my issue:
Hi,

I have setup akeeba backup to run automatically as per a pre-defined schedule.

The post processing is setup to copy the files to google drive and then delete the older file from google drive as per the quota count.

I have setup the cron jobs correctly and they have been running well.

Since some time ago, sometimes the process which gets started by the cron job, never gets completed and I have to login to the shell prompt and manually kill the akeeba process. When this happens, the backup is successful locally, the transfer to goole drive is also successful. HOWEVER, the deletion of the older file from google drive as per the quota retention count - that does not happen. I think the process gets stuck somewhere. I have attached the log file of one such process that got stuck. The log file is zipped.

In this particular case where I am sending the log file - the backup process is completed locally, but the file is not got transferred to google drive. I must have killed the process before file transfer got completed.

Please help. Not urgent, but still important, because it's kind of painful to manually kill the process everynow and then.

Your log file is not coming trough, can you please try again? If it won't post, just put it on a file service, Dropbox, Google Drive, etc. and give me the URL, I'll go get it. Thank you.

Oh boy.. I didn't see that the log file did not get uploaded and then deleted the same..

My bad.

Please keep this pending. I am sure this will happen again with a day or two..

Will upload log file when it happens again.

If it happens that this ticket auto closes before you get another failed log, just open a new ticket. There's no history on this one so nothing to worry about.

I have uploaded a zip file to google drive with share link - https://drive.google.com/open?id=1dkuJ8nfxMasB2c-1oryVvpP3lv7CHBaH

The zip file contains 2 logs - One log is for a properly completed backup process and the other is for the process stuck. You will know from the name given to the log file.

Also trying to attach the zip file here..

It may have to do with the archive sizes. The failed one is 250Mb, the successful one is 71Mb.

Please change the part size for the archive for Profile #2 to 50Mb. You can find that setting under Configuration, Archiver Engine, Configuration. It is currently set to 2000. That will give you a backup archive with 5-6 parts. All of the parts are necessary to restore the backup. If this works, we can play with the part size a little bit to see if a larger part will work.

I implemented the suggested configuration.

Seems to work so far. Is there anything else that i can check/do?

Your upload to Google Drive uses a "chunk" method. The full archive is broken into 5Mb chunks, the chunk is uploaded to Google, when they are all uploaded, Google reassembles all the chunks into a single archive file. This process is designed to work around the PHP time limit. The full fine must be uploaded within the time limit. But in this case, only a single chunk needs to be uploaded, then the time limit starts over and the next chunk goes up.

As Davide explained it to me, the full archive file needs to be loaded into memory in order to create the chunks on the server side. That's where the problem is arising. We're running out of memory for the full archive to be loaded and chunked.

The solution is to break the archive into parts. "Parts" break your backup archive into pieces, all of the pieces are necessary to extract the archive. Where parts and chunks differ is that the parts don't get reassembled, your archive remains multiple parts.

We know now that a 250Mb file size archive does not work. A 50Mb archive part size does work. How big can we make the part size and still allow the part to be loaded into memory and chunked for upload? That I can't answer. You will need to play around with the part size until the upload starts to fail again, then back it off some. We don't want to be too aggressive with the part size, we want the upload to be consistent. 50Mb works, next try 100Mb, then 150Mb.

This ticket has been automatically closed. All tickets which have been inactive for a long time are automatically closed. If you believe that this ticket was closed in error, please contact us.