Akeeba Backup for Joomla!

Thursday, 01 February 2018 03:04 CST

guzabi

Hello,

I have a warning when performing a backup task with post-processing set to upload to Amazon S3. I have a warning saying :

Warnings
0 :: Akeeba\Engine\Postproc\Connector\S3v4\Connector::putObject(): [0] Access Denied Debug info: SimpleXMLElement Object ( [Code] => AccessDenied [Message] => Access Denied [RequestId] => 125CFEA9E7C5AFF1 [HostId] => nP48BCcw8EirUc3vuboXrWncX3uwAGFfoeVyg++xDeGquQ7wc1qxX/O4OmjSgRkshwqymzLUvKM= )
Failed to upload kickstart.php

The backup seems fine, though. But the warning is worrying.
Is this a known bug? Should I worry about this ?

Thursday, 01 February 2018 04:53 CST

tampe125

Hello,

it seems that the S3 user you are using hasn't the correct permissions to upload the backup archive.
You should review your user permissions and double check that he has write permissions on the bucket you're uploading to.

Davide Tampellini

Developer and Support Staff

🇮🇹Italian: native 🇬🇧English: good • 🕐 My time zone is Europe / Rome (UTC +1)
Please keep in mind my timezone and cultural differences when reading my replies. Thank you!

Thursday, 01 February 2018 05:10 CST

guzabi

Thanks for the answer. This is what I thought, however it's not the case :

1. The user and his rights are unchanged since months, and the same user is used on other websites, and the upload goes very well, without warnings ;

2. The JPS is indeed uploaded. There's 1.1Gb of data on S3, uploaded at the right time…

Please note that the user is write-only (for security reasons) and that the archive is huge, here (1.1Gb+). The other websites uploading to S3 generate much smaller archives (100 ~ 200 Mb). Maybe that's related, I don't know…

Edited by guzabi on 2018-02-01 05:12 CST

Thursday, 01 February 2018 05:28 CST

tampe125

If you try again, is it uploaded?
Sometimes Amazon fails on his own and throws cryptic errors.

Davide Tampellini

Developer and Support Staff

🇮🇹Italian: native 🇬🇧English: good • 🕐 My time zone is Europe / Rome (UTC +1)
Please keep in mind my timezone and cultural differences when reading my replies. Thank you!

Thursday, 01 February 2018 05:30 CST

guzabi

I saw the warning this morning on the backup notification e-mail (the backup is triggered by a cron task). I was worried, so I logged on and manually started the backup (same profile). I also got the warning. Then I logged on to my AWS console and checked the bucket, and I saw that both archives (from the automatic backup, during the night, and from my manual backup, this morning) were there, with the right modification time and size…

Thursday, 01 February 2018 07:11 CST

tampe125

Yeah, sometimes it happens: Amazon S3 panics on its own and throws out an error. This is not the first time I see something like that.
Usually those are just temporary issues. I'll leave this ticket open, let me know if that still happens in following days or it was just a one time issue.

Davide Tampellini

Developer and Support Staff

🇮🇹Italian: native 🇬🇧English: good • 🕐 My time zone is Europe / Rome (UTC +1)
Please keep in mind my timezone and cultural differences when reading my replies. Thank you!

Sunday, 18 February 2018 14:45 CST

guzabi

Hello,
Coming back on this ticket. I now see that I have a problem with local backup, but with a "slient fail" this time. The notification e-mail simply stops after a while, not displaying a conclusion or error notification (log is attached). The backup fails, no archive is generated. Manual backup with same profile works, though.
This is my only website with errors, and this is the same every time : cron backup fails, manual backup succeeds (using identical profiles). Is there something different when working with cron backups? What's the issue, here? The website is particularly huge (archives are ±1.30 Gb each) but if that was an issue, errors should happen woth manual backup as well, no?
Please, please help…

Monday, 19 February 2018 02:41 CST

tampe125

It seems your log didn't come trough, however I have an hunch: most likely your CRON job has an hard limit on a fixed amount of minutes.
On manual backups you don't have such problem, since we're requesting the same page several times, resetting the timer. On CLI backups, instead, everything happens on the same request, thus the problem.
The only solution is to trigger them using frontend backups, further info here

Davide Tampellini

Developer and Support Staff

🇮🇹Italian: native 🇬🇧English: good • 🕐 My time zone is Europe / Rome (UTC +1)
Please keep in mind my timezone and cultural differences when reading my replies. Thank you!

Wednesday, 21 March 2018 18:17 CDT

System Task

system

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.

Support