Experience: PBS: CIFS Storage

hellfire

Renowned Member
Aug 17, 2016
83
51
83
47
Hi,

just a short note. For a customer I had setup PBS with Storage on WebDAV CIFS. CIFS is a quite slow Network storage, so in terms of choice of technology it's very poor. Nevertheless, a big amount of storage space is available here for a very cheap price, so I tried it. Local Disk Storage is more expensive by magnitudes. I have two Proxmox VE Hosts, each serving only one big virtual machine and it works quite well.

Recovery is at a quite acceptable speed. Nothing dead slow.

Regards,
h.

Edit: I noticed that it was not WebDAV - which would be even worse - but CIFS.
 
Last edited:
  • Like
Reactions: noel.
What about GC and verify job performance? These are usually what takes the most time, because of all the small random IO.
 
  • Like
Reactions: hellfire
Good Question!

The backup jobs run daily.

Here are some durations in format "Start-Time(YYYY-MM-DDTHH:MM:SS) Duration(HH:MM:SS)" from the logs.

Here are the garbage job times - they're rising through the roof.

Code:
2022-11-08T20:18:40 00:56:33
2022-11-09T00:00:00 00:37:13
2022-11-10T00:00:00 00:53:50
2022-11-11T00:00:00 00:54:20
2022-11-12T00:00:00 00:56:18
2022-11-13T00:00:00 00:55:19
2022-11-14T00:00:00 00:56:47
2022-11-15T00:00:00 00:54:12
2022-11-16T00:00:00 01:52:29
2022-11-17T00:00:00 02:10:38
2022-11-18T00:00:00 02:08:33
2022-11-19T00:00:00 03:11:16
2022-11-20T00:00:00 02:53:06
2022-11-21T00:00:00 03:37:19
2022-11-22T00:00:00 05:08:43
2022-11-23T00:00:00 04:04:12
2022-11-24T00:00:00 07:49:25
2022-11-25T00:00:00 08:53:33
2022-11-26T00:00:00 11:25:27
2022-11-27T00:00:00 13:06:06
2022-11-28T00:00:00 16:37:18
2022-11-29T00:00:00 15:19:46
2022-11-30T00:00:00 15:17:45
2022-12-01T00:00:00 14:57:24
2022-12-02T00:00:00 16:24:35
2022-12-03T00:00:00 17:47:17
2022-12-04T00:00:00 10:30:34
2022-12-05T00:00:00 12:53:26
2022-12-06T00:00:00 03:28:47
2022-12-07T00:00:00 04:36:54
2022-12-08T00:00:00 05:57:55
2022-12-09T00:00:00 14:08:22
2022-12-10T00:00:00 14:41:06
2022-12-11T00:00:00 00:00:00
2022-12-12T00:00:00 14:53:01
2022-12-13T00:00:00 00:00:00

Here are the verification Job times - quite normal(daily runs):

Code:
2022-11-12T09:00:00 00:00:00
2022-11-15T15:11:31 00:21:31
2022-11-16T13:43:05 00:21:13
2022-11-19T09:00:00 00:40:58
2022-11-21T21:00:00 02:54:06
2022-11-22T21:00:00 00:56:53
2022-11-23T21:00:00 01:29:48
2022-11-24T21:00:00 02:02:43
2022-11-25T21:00:00 02:02:14
2022-11-26T21:00:00 01:49:39
2022-11-27T21:00:00 01:53:03
2022-11-28T21:00:00 01:51:30
2022-11-29T21:00:00 02:29:35
2022-11-30T21:00:00 03:49:12
2022-12-01T21:00:00 03:57:08
2022-12-02T21:00:00 03:58:35
2022-12-03T21:00:00 02:32:16
2022-12-04T21:00:00 02:41:28
2022-12-05T21:00:00 02:38:45
2022-12-06T21:00:00 03:45:33
2022-12-07T21:00:00 03:44:22
2022-12-08T21:00:00 03:58:38
2022-12-09T21:00:00 04:17:22
2022-12-10T21:00:00 02:16:35
2022-12-11T21:00:00 05:18:00
2022-12-12T21:00:00 02:06:33

Here are the BACKUP Job times for vm #1 (webserver)

Code:
2022-11-08T18:05:28 00:00:01
2022-11-08T18:08:50 00:00:00
2022-11-08T18:14:14 00:00:34
2022-11-08T18:15:38 00:00:01
2022-11-21T19:41:35 00:51:51
2022-11-21T20:42:27 00:00:32
2022-11-21T20:43:09 00:00:03
2022-11-21T22:00:04 00:00:13
2022-11-22T22:00:05 00:06:30
2022-11-23T22:00:03 00:00:49
2022-11-24T22:00:07 00:06:30
2022-11-25T22:00:05 00:06:05
2022-11-26T22:00:05 00:09:00
2022-11-27T22:00:03 00:10:10
2022-11-28T22:00:05 00:34:34
2022-11-29T22:00:09 01:39:46
2022-11-30T22:00:05 00:55:32
2022-12-01T22:00:08 00:17:26
2022-12-02T22:00:04 00:22:38
2022-12-03T22:00:05 00:11:40
2022-12-04T22:00:06 00:13:49
2022-12-05T22:00:05 00:16:26
2022-12-06T22:00:05 00:29:57
2022-12-07T22:00:07 00:27:46
2022-12-08T22:00:07 00:30:28
2022-12-09T15:54:43 00:13:50
2022-12-09T22:00:04 00:09:51
2022-12-10T22:00:05 00:24:28
2022-12-11T22:00:07 00:33:52
2022-12-13T03:00:05 00:24:25

Here are the BACKUP Job times for vm #2 (DB-Server)

Code:
2022-11-15T14:26:25 00:42:09
2022-11-15T15:10:45 00:00:08
2022-11-15T15:12:49 00:00:04
2022-11-15T21:00:03 00:00:37
2022-11-16T21:00:01 00:29:00
2022-11-17T21:00:03 00:25:25
2022-11-18T21:00:01 00:28:46
2022-11-19T21:00:02 00:22:59
2022-11-20T21:00:04 00:31:55
2022-11-21T21:00:03 00:11:22
2022-11-22T21:00:02 00:27:52
2022-11-23T21:00:01 00:38:40
2022-11-24T21:00:03 00:43:47
2022-11-25T21:00:03 00:42:00
2022-11-26T21:00:03 00:36:50
2022-11-27T21:00:03 00:36:57
2022-11-28T21:00:02 00:32:47
2022-11-29T21:00:03 00:42:47
2022-11-30T21:00:04 01:28:20
2022-12-01T21:00:03 00:50:00
2022-12-02T21:00:03 01:30:34
2022-12-03T21:00:01 00:30:22
2022-12-04T21:00:02 00:36:36
2022-12-05T21:00:01 00:47:40
2022-12-06T21:00:04 01:25:30
2022-12-07T21:00:03 01:44:33
2022-12-08T21:00:04 01:22:42
2022-12-09T21:00:03 00:55:32
2022-12-10T21:00:02 05:23:48
2022-12-11T21:00:05 01:24:34
2022-12-12T21:00:03 05:06:59
 
Last edited:
there is overlaps in jobs, so there is a slowdown because destination disks are busy.
imo, daily GC is an over kill.
here, it's set on each sunday.
verify job is set sunday too after the GC.
 
  • Like
Reactions: ITT
Duration of the jobs is fine for me now. And by the way: I further impose load on the system by using BackupPC (=File based Network Backup Solution) on the system and the same storage too.

Garbage Collector (run set to weekly):

screenshot_garbage_collector.png

Backup Durations: #100 = web-server #102 = db-server
screenshot_backup.png
 
Last edited: