syslog question

[RobFantini]

at pbs > administration > syslog , there are many lines like the following. 95%+ of the log:

Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:07 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:17 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:17 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:17 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:17 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'
Jul 28 22:13:17 s022 proxmox-backup-api[8259]: successful auth for user 'pbs-user@pbs'

So I tried using this to filter out:
/etc/rsyslog.d/pve-local.conf :

if ($programname == 'proxmox-backup-api' ) then stop
:msg, contains, "successful.auth.for.user" stop

that worked for preventing those logs going to /var/log/syslog

however the useless lines still show up at pbs>admin>syslog .

Is there a way to filter those from cluttering up pbs syslog?

 

[fabian]

filed https://bugzilla.proxmox.com/show_bug.cgi?id=2903 with the likely culprit. journald does not offer such a filtering option AFAIK

 

[RobFantini]

OK thanks for filing the report.

you know things are getting good with PBS when these not so important items get reported. overall PBS is working great.

 

[Apollon77]

Also see https://forum.proxmox.com/threads/high-load-because-of-pve-storage-checks.73756/#post-329300 ... in fact yes it is ok for PBS for me (beside the log spam) ... but I have more probs with the pve performance as soon as I activate a pbs storage there

 

[RobFantini]

We are not showing high cpu or memory usage at our PBS system.

On our 7 node cluster, each node backs up a different day . so we do not have concurrent backups. perhaps concurrent backups or low

hardware info for our pbs:

lshw -short -quiet|less
system         X10SLH-N6-ST031
processor      Intel(R) Xeon(R) CPU E3-1246 v3 @ 3.50GHz

8GB Memory

zpool, note we have cache and log :

zpool status tank2
  pool: tank2
 state: ONLINE
  scan: resilvered 52.1G in 0 days 00:20:39 with 0 errors on Wed Jul 15 12:30:51 2020
config:

        NAME                                            STATE     READ WRITE CKSUM
        tank2                                           ONLINE       0     0     0
          mirror-0                                      ONLINE       0     0     0
            ata-Hitachi_HDS5C3030ALA630_MJ1311YNG6Y6LA  ONLINE       0     0     0
            ata-Hitachi_HDS5C3030ALA630_MJ1311YNG6Y93A  ONLINE       0     0     0
        logs
          wwn-0x55cd2e404b65b8ab-part1                  ONLINE       0     0     0
        cache
          wwn-0x55cd2e404b65b8ab-part2                  ONLINE       0     0     0

the system never uses swap. however we are soon adding memory as zfs systems usually do better with more..

 

[Apollon77]

@RobFantini Do not get me wring: Not the backup itself uses up the resources ... it is just when I add the pbs to PVE as storage ... then it starts

 

[RobFantini]

@RobFantini Do not get me wring: Not the backup itself uses up the resources ... it is just when I add the pbs to PVE as storage ... then it starts

OK .