scalix server perfomance issue ad secundam

Discuss the Scalix Server software

Moderators: ScalixSupport, admin

stephan.klein
Posts: 68
Joined: Thu Feb 22, 2007 1:10 pm
Location: Germany

scalix server perfomance issue ad secundam

Postby stephan.klein » Mon Apr 02, 2007 4:12 pm

Hi,

after I put Scalix (latest release) in production for the first ten users, I discovered performance problems with the imap daemon.

If one user deletes or moves messages (about 50 up) via Thunderbird (1.5, 2 beta), the imap subsystem is nearly unusable for the other users.

When moving 80 messages to trash 'top' shows:

Code: Select all

top - 22:03:47 up 1 day,  8:50,  4 users,  load average: 0.56, 0.51, 0.84
Tasks: 201 total,   1 running, 200 sleeping,   0 stopped,   0 zombie
Cpu0  :  0.7%us,  1.7%sy,  0.0%ni, 75.6%id, 22.1%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu1  :  0.7%us,  1.7%sy,  0.0%ni, 72.3%id, 25.4%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu2  :  0.0%us,  1.0%sy,  0.0%ni, 70.3%id, 28.7%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu3  :  0.3%us,  0.0%sy,  0.0%ni, 75.5%id, 24.2%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   1554928k total,  1508432k used,    46496k free,   113704k buffers
Swap:   803240k total,      472k used,   802768k free,   706276k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
16180 60536     18   0 43796  24m 5080 D    3  1.6   9:36.39 in.imap41d
  786 root      15   0  2356 1200  856 R    1  0.1   0:01.05 top
24641 root      16   0  2356 1200  856 S    1  0.1   0:58.78 top
...



A 'ps waux |grep imap shows

Code: Select all

mail7:/home/stephan# ps waux |grep imap
60536      528  0.0  0.3  23488  6068 ?        Ss   21:54   0:00 in.imap41d -d
60536      991  0.1  0.3  23552  6104 ?        Ss   22:02   0:00 in.imap41d -d
60536     1009  0.1  0.3  23148  5644 ?        Ss   22:02   0:00 in.imap41d -d
60536     1032  0.3  0.4  23884  6804 ?        Ss   22:03   0:00 in.imap41d -d
root      1421  0.0  0.0   1648   564 pts/2    S+   22:05   0:00 grep imap
60536     3285  0.0  0.4  23972  7032 ?        Ss   16:29   0:01 in.imap41d -d
root      4512  0.0  0.1  16004  1836 ?        Ss   Apr01   0:01 in.imap41d -d
60539    13629  0.0  0.3  23148  5776 ?        Ss   19:14   0:00 in.imap41d -d
60539    13630  0.0  0.4  26028  6676 ?        S    19:14   0:00 in.imap41d -d
60536    16180  1.3  1.5  43796 24856 ?        D    10:25   9:37 in.imap41d -d
60534    17036  0.0  0.3  22588  5356 ?        Ss   14:44   0:00 in.imap41d -d
60534    17037  0.0  0.4  26184  6276 ?        S    14:44   0:00 in.imap41d -d
60540    27649  0.0  0.4  25964  6356 ?        S    21:23   0:00 in.imap41d -d
60536    28033  0.0  0.4  23380  6264 ?        Ss   21:33   0:00 in.imap41d -d
60540    28673  0.0  0.3  23116  5940 ?        Ss   21:36   0:00 in.imap41d -d


The filesystem is ext3, I installed an additional drive with ext2 for the temp directory.

Any hints are welcome, kind regards

Stephan

craig

Postby craig » Mon Apr 02, 2007 5:16 pm

I see all the Imap connections, but I see no load on the server.

stephan.klein
Posts: 68
Joined: Thu Feb 22, 2007 1:10 pm
Location: Germany

Postby stephan.klein » Mon Apr 02, 2007 5:35 pm

That's my problem - I have no idea where to tweak the machine. An other thread here on the forum ('heavy imap use is killing the server') mentioned the I/O subsystem, but an average I/O wait of 30% per CPU does not seem too much for me.

Does scalix take any benefit of multiple CPUs?

Stephan


Return to “Scalix Server”



Who is online

Users browsing this forum: No registered users and 10 guests

cron