Facebook
TwitterTheEternalNocturnal
New Pleskian
Hi all,
I am fighting an ongoing backup issue on four Linux Plesk 11.0.9 servers (recently upgraded from Plesk 9.5.4, which also had the same problems) where the backups run all day instead of at their scheduled times.
All the backups are set to run between midnight and 6am within the Plesk GUI, but what seems to happen is that every day, around 10-11am, the backups all kick off at once causing a high server load. The server becomes unresponsive until all these backup processes have subsided, whereupon the server returns to normal. Whilst this is happening, top -c reveals multiple plesk_agent_manager processes running.
If I kill these processes, the server starts to return to normal... except the processes then start themselves back up again! The only way to get through this is to let Plesk do what it wants to do, which takes anything from ten minutes to an hour.
This doesn't happen every day on every server, but it seems to happen to at least one server nearly every day!
I have tried many many things, but recently I've tried the following to control when the backups could run -
1/ First I amended the scheduled times inside /etc/cron.d/plesk-backup-manager. By default it's scheduled to run backupmng at 8,23,38 and 53 mins past every hour. So I restricted the hours it could run, from midnight to 7am. No effect - all backup jobs still kick off at once as described above.
2/ So on one server I took the step of setting up a cron job to rename /usr/local/psa/admin/bin/plesk_agent_manager at 7am, so the backups simply cannot run, and rename it back again at midnight. This prevented the backups running... they stopped completely. Despite several reboots the backups simple failed to start back up again, until about three days later when they unexpectedly kicked back into life again causing the same problem. At this point I disabled them and instead switched to taking a daily full server backup (using /usr/local/psa/bin/pleskbackup server, which incidentally runs weekly on all eight of my Linux Plesk servers generally without any issue! But this is no solution long-term as end users need access to their individual backups).
Here's an example from one of the servers. This is simply a list of "backup_time" and "last" from Backups Scheduled -
02:00:00 2013-06-13 10:27:55
03:00:00 2013-06-13 09:57:38
03:30:00 2013-06-13 10:06:08
04:00:00 2013-06-13 10:12:35
04:30:00 2013-06-13 09:59:42
05:00:00 2013-06-13 09:59:52
05:30:00 2013-06-13 10:11:32
01:30:00 2013-06-13 01:32:10
06:00:00 2013-06-13 10:00:57
03:00:00 2013-06-13 10:21:06
04:15:00 2013-06-13 10:16:44
01:00:00 2013-06-13 01:10:25
01:20:00 2013-06-13 01:42:10
01:00:00 2013-06-13 01:02:19
02:00:00 2013-06-13 11:40:25
04:10:00 2013-06-13 09:58:43
03:05:00 2013-06-13 10:09:03
02:20:00 2013-06-13 10:25:27
01:30:00 2013-06-13 01:56:19
02:35:00 2013-06-13 10:01:49
04:30:00 2013-06-13 10:10:29
03:30:00 2013-06-13 09:58:42
02:23:00 2013-06-13 09:59:42
02:15:00 2013-06-13 10:01:52
03:00:00 2013-06-13 10:38:13
02:00:00 2013-06-13 10:32:07
All the other backup settings are identicle for each backup -
obj_type = domain
repository = ftp
period = 86400
active = true
processed = false
rotation = 7
split_size = 0
suspend = false
with_content = true
backup_day = 0
content_type = backup_content_all_at_domain
So my question is.... why do the backups not run at the scheduled times, but instead run during the day?!!
If anybody has any advice or can point me toward any relevant articles, or maybe explain how the Plesk backup scheduler works (or doesn't!
, that would be very much appreciated!
Many thanks.
TEN.
I am fighting an ongoing backup issue on four Linux Plesk 11.0.9 servers (recently upgraded from Plesk 9.5.4, which also had the same problems) where the backups run all day instead of at their scheduled times.
All the backups are set to run between midnight and 6am within the Plesk GUI, but what seems to happen is that every day, around 10-11am, the backups all kick off at once causing a high server load. The server becomes unresponsive until all these backup processes have subsided, whereupon the server returns to normal. Whilst this is happening, top -c reveals multiple plesk_agent_manager processes running.
If I kill these processes, the server starts to return to normal... except the processes then start themselves back up again! The only way to get through this is to let Plesk do what it wants to do, which takes anything from ten minutes to an hour.
This doesn't happen every day on every server, but it seems to happen to at least one server nearly every day!
I have tried many many things, but recently I've tried the following to control when the backups could run -
1/ First I amended the scheduled times inside /etc/cron.d/plesk-backup-manager. By default it's scheduled to run backupmng at 8,23,38 and 53 mins past every hour. So I restricted the hours it could run, from midnight to 7am. No effect - all backup jobs still kick off at once as described above.
2/ So on one server I took the step of setting up a cron job to rename /usr/local/psa/admin/bin/plesk_agent_manager at 7am, so the backups simply cannot run, and rename it back again at midnight. This prevented the backups running... they stopped completely. Despite several reboots the backups simple failed to start back up again, until about three days later when they unexpectedly kicked back into life again causing the same problem. At this point I disabled them and instead switched to taking a daily full server backup (using /usr/local/psa/bin/pleskbackup server, which incidentally runs weekly on all eight of my Linux Plesk servers generally without any issue! But this is no solution long-term as end users need access to their individual backups).
Here's an example from one of the servers. This is simply a list of "backup_time" and "last" from Backups Scheduled -
02:00:00 2013-06-13 10:27:55
03:00:00 2013-06-13 09:57:38
03:30:00 2013-06-13 10:06:08
04:00:00 2013-06-13 10:12:35
04:30:00 2013-06-13 09:59:42
05:00:00 2013-06-13 09:59:52
05:30:00 2013-06-13 10:11:32
01:30:00 2013-06-13 01:32:10
06:00:00 2013-06-13 10:00:57
03:00:00 2013-06-13 10:21:06
04:15:00 2013-06-13 10:16:44
01:00:00 2013-06-13 01:10:25
01:20:00 2013-06-13 01:42:10
01:00:00 2013-06-13 01:02:19
02:00:00 2013-06-13 11:40:25
04:10:00 2013-06-13 09:58:43
03:05:00 2013-06-13 10:09:03
02:20:00 2013-06-13 10:25:27
01:30:00 2013-06-13 01:56:19
02:35:00 2013-06-13 10:01:49
04:30:00 2013-06-13 10:10:29
03:30:00 2013-06-13 09:58:42
02:23:00 2013-06-13 09:59:42
02:15:00 2013-06-13 10:01:52
03:00:00 2013-06-13 10:38:13
02:00:00 2013-06-13 10:32:07
All the other backup settings are identicle for each backup -
obj_type = domain
repository = ftp
period = 86400
active = true
processed = false
rotation = 7
split_size = 0
suspend = false
with_content = true
backup_day = 0
content_type = backup_content_all_at_domain
So my question is.... why do the backups not run at the scheduled times, but instead run during the day?!!
If anybody has any advice or can point me toward any relevant articles, or maybe explain how the Plesk backup scheduler works (or doesn't!
, that would be very much appreciated! Many thanks.
TEN.
