----- Original Message -----

From: Henrik Woffinden

To: surgemail-list@netwinsite.com

Sent: Tuesday, March 01, 2011 2:00 AM

Subject: Re: [SurgeMail List] Slow: Use task manager or top to check system load, check for cpu/disk/memoryusage

I do have NTP daemon running on the server, but the OS system log doesn't indicate time changes larger than half a second.

Surgemails mail.err says:
26 21:51:02.00:Err:-1222202512: selfmonitor: cpu seems busy slowed by factor 2.3
27 00:23:42.00:Err:-1221768336: selfmonitor: thread lockup occured for 77 seconds
27 00:42:27.00:Err:-1221768336: selfmonitor: thread lockup occured for 11 seconds
27 13:01:46.00:Err:-1221768336: selfmonitor: thread lockup occured for 453 seconds
27 13:14:46.00:Err:-1221768336: selfmonitor: thread lockup occured for 8 seconds
28 14:34:00.00:Err:-1221768336: selfmonitor: thread lockup occured for 923 seconds
28 17:05:52.00:Err:-1221768336: selfmonitor: cpu seems busy slowed by factor 2.2
 1 00:00:06.00:Err:-1221768336: selfmonitor: cpu seems busy slowed by factor 2.3

Of course, you are the expert here, but I find it hard to believe that Nagios has nothing to say with a lockup for 923 seconds. I was connected to the server at that time, and I didn't notice any lockups.

I do have a mirror slave server which internal clock is rather slow, so NTP daemon often adjusts system time there, but I guess that those adjustments wouldn't appear in mail.err on the master server. Am I right?

/Henrik

On 28-02-2011 23:32, Support ChrisP wrote:

That message is triggered by this code:

last = time(NULL);

lib_msleep(100); 

if (time(NULL)-last>7) {

So, basically, a thread goes to sleep for 1/10th of a second, and if it doesn't get any 'run' time for 7 seconds after that (which in computer terms is an unusually long time), then it generates that warning.

This probably means 'something' on your system is locking up the whole server for a few seconds once every few hours...  

If you examine mail.err you should see the actual delay that surgemail is noticing... 

       emsg("selfmonitor: thread lockup occured for %d seconds\n",time(NULL)-last);

One other thing that can cause this is if you have a cron job that changes the time on the system :-) 

Also if you check the time of those log entries in mail.err maybe you will spot a pattern that might explain the issue  

And check your syslog entries for the same times in case that gives a clue.

If the delay is relatively short and no other symptoms appear then you can safely ignore these warning emails.  

      

      ChrisP.

On Tuesday 01/03/2011 at 3:07 am, Henrik Woffinden wrote:

Hi Support,

After upgrading my server from Fedora 10 to Fedora 14, this weekend I get this message every few hours:
------- cut ---------
Important report from SurgeMail woffinden.co.uk host=delias.woffinden.co.uk - see details below

Slow: Use task manager or top to check system load, check for cpu/disk/memory usage
-------- cut --------

On the same server I also run Nagios, an here everything is fine:

26-02-2011 00:00:00	26-02-2011 15:32:07	0d 15h 32m 7s	SERVICE OK (HARD)	OK - load average: 0.00, 0.02, 0.00
27-02-2011 00:00:00	28-02-2011 00:00:00	1d 0h 0m 0s	SERVICE OK (HARD)	OK - load average: 0.00, 0.01, 0.05
28-02-2011 00:00:00	28-02-2011 14:51:23	0d 14h 51m 23s+	SERVICE OK (HARD)	OK - load average: 0.00, 0.02, 0.05

What CPU/Memory/disk levels triggers SurgeMail?

The Partition where SurgeMail resides has 135GB free (31% free)

Hardware is 1 Core 2.8GHz HT with 4GB RAM.
As you can see in below SurgeMail info, it's not a busy server.

Version: Linux 5.2a-1 Dec 20 2010 20:21:06 (Latest surgemail version installed)
Uptime : 1 days 16:43:41        Server time: Mon Feb 28 15:01:59 2011

Authent: Lookups 195, cache 51592, no such user 116, cache size 114, speed 0.35/s
Smite  : 36.7% 4026/10974
Avast  : (beta) see advanced page for details
SSL/TLS: Allow[*], OpenSSL 0.9.7l 28 Sep 2006

MESSAGES                           avg(/s)          MIRRORING
   In: 11450       Sent: 263        0.00               Add:  36660  (2.2/s) live=0
  Out: 11514     Stored: 11144      0.08              Sent:  19447  (1.2/s)
Queue: 6        Bounced: 5          0.00             Queue:      0k  0.0%
                Dropped: 83         0.00            Config: Feb 27 21:29:18

CONNECTIONS                      [Recent: last 32s]     DISK FREE
        now/peak  denied   avg   [avg   cpu  bandw]
      (concurrent) (tot)  (/s)   [/s]   [%] [byte/s]       Home: 140,887,917k (/usr/local/surgemail)
Other:    0/2         0   0.00   0.00    0%    0.0k
 SMTP:    0/15        0   0.11   0.12    0%    0.1k
  POP:    0/4         0   0.00   0.00    0%    0.0k
  WEB:    1/6         0   0.01   0.41    0%    2.2k
 IMAP:   30/38        0   0.28   0.47    0%    1.8k
 SEND:    0/4         0   0.00   0.00    0%    0.0k
SURGEWEB:    0/0         0   0.00   0.00    0%    0.0k
   SW:    0/0         0   0.00   0.00    0%    0.0k
Total:   31/46        0   0.40   1.00    0%    4.1k

[DNS]                Ok(avge)
      127.0.0.1:  12881(0.1s)
  xx.xx.xx.xx:  12880(0.1s)
  xx.xx.xx.xx:      0(0.0s)

USERS
Total (ever logged into POP/IMAP):         68
  Migrated:          0 this month
       New:          0 this month
 SurgePlus:         25

--
Med venlig hilsen / Best regards,

Henrik Woffinden