Issue /usr/local/psa/handlers/spool full of old emails

[Frostbolt]

I noticed our harddisc was almost full today, and upon investigating I found /usr/local/psa/handlers/spool full of old emails, literally hundreds of thousands.

I found this topic: /usr/local/psa/handlers/spool full of e-mails

It states the issue was resolved in Plesk 12.0 but I'm running Plesk v17.8.11 and the emails aren't getting deleted for us.

Edit: I see I've posted this question in the wrong forum, sorry! Can anyone move it please?

 

[Ales]

Are you seeing any other errors on the server, anything in the /var/log/messages (CentOS) a.k.a. /var/log/syslog (Debian and derivatives)?

There might be an error somewhere in the mail path that causes these files to linger instead of being removed.

 

[Ales]

Refer to this thread too, it might be helpful: DMARC Changes in MU#51 Introduce New Segfault Error (PPPM-10547).

 

[Frostbolt]

Yeah, I'm seeing a lot of:

Nov 12 05:04:58 _ xinetd[1298]: START: ftp pid=32586 from=::ffff:1.56.233.97
Nov 12 05:04:59 _ xinetd[1298]: EXIT: ftp status=0 pid=32586 duration=1(sec)
Nov 12 05:04:59 _ xinetd[1298]: START: ftp pid=32590 from=::ffff:1.56.233.97
Nov 12 05:05:02 _ xinetd[1298]: EXIT: ftp status=0 pid=32590 duration=3(sec)

and

Nov 12 05:05:02 _ kernel: show_signal_msg: 80 callbacks suppressed
Nov 12 05:05:02 _ kernel: dk_sign[32638]: segfault at 8 ip 00000000004028d3 sp 00007fff83988200 error 4 in dk_sign[400000+e000]
Nov 12 05:05:02 _ kernel: dk_sign[32647]: segfault at 8 ip 00000000004028d3 sp 00007fff40c7d510 error 4 in dk_sign[400000+e000]
Nov 12 05:05:02 _ kernel: dk_sign[32656]: segfault at 8 ip 00000000004028d3 sp 00007fff7ed42540 error 4 in dk_sign[400000+e000]
Nov 12 05:05:02 _ kernel: dk_sign[32665]: segfault at 8 ip 00000000004028d3 sp 00007fff01d18fb0 error 4 in dk_sign[400000+e000]

 

[Ales]

The xinetd entries are normal, or, better said, they might mean that someone is brute forcing the FTP, but they aren't connected to this issue. The segmentation faults are directly connected, though.

They basically mean that DKIM is crashing and leaving those messages behind. The issue is likely different than the one I've linked above (PPPM-10547), so I'd advise you to fill in a separate bug report.