Hi,
>________________________________________
>De : [EMAIL PROTECTED] [EMAIL PROTECTED] de la part de Victor Duchovni [EMAIL
>PROTECTED]
>Date d'envoi : mardi 28 octobre 2008 20:27
>À : postfix-users@postfix.org
>Objet : Re: RE : Big incoming queue, slow qmgr, idle system.
>
>On Tue, Oct 28, 2008 at 07:50:51PM +0100, Francis SOUYRI wrote:
>
>> >> Time to time we have lot of mail in the incoming queue, "no" mail in the
>> >> active queue, the "qmgr "is very" slow and the server is very idle. We
>> >> tested the antispam/virus to see if there is a problem but the response
>> >> is very good.
>> >
>> >How often does the qmgr(8) process id change?
>>
>> No change, only when I restart postfix.
>
>And how often do you do that?
For this day 15 times, all are postfix restarts, after parameters changes to
see what appended (activate/deactivate qmgr verbose) or when there is too many
mails in the incoming (to reduce the smtpd from 100 to 40 in the master.cf)
Oct 28 11:18:42 vador postfix/qmgr[11389]: dict_eval: const mail
Oct 28 11:25:53 vador postfix/qmgr[16544]: dict_eval: const mail
Oct 28 11:27:39 vador postfix/qmgr[17347]: dict_eval: const mail
Oct 28 11:59:27 vador postfix/qmgr[6309]: dict_eval: const mail
Oct 28 12:01:04 vador postfix/qmgr[7741]: C74BB481BA: from=<>, size=4328,
nrcpt=1 (queue active)
Oct 28 12:05:08 vador postfix/qmgr[12042]: dict_eval: const mail
Oct 28 12:11:19 vador postfix/qmgr[15251]: B3A02484E0: from=<[EMAIL
PROTECTED]>, size=8580, nrcpt=1 (queue active)
Oct 28 12:46:32 vador postfix/qmgr[16259]: 527E348896: from=<[EMAIL
PROTECTED]>, size=3218, nrcpt=1 (queue active)
Oct 28 15:03:13 vador postfix/qmgr[8000]: D4CC5485BF: from=<[EMAIL PROTECTED]>,
size=10884, nrcpt=1 (queue active)
Oct 28 15:04:20 vador postfix/qmgr[9029]: 7CBD748169: from=<[EMAIL PROTECTED]>,
size=3611, nrcpt=1 (queue active)
Oct 28 15:05:01 vador postfix/qmgr[9536]: 74406481F6: from=<[EMAIL PROTECTED]>,
size=2868, nrcpt=10 (queue active)
Oct 28 15:05:42 vador postfix/qmgr[10173]: AE4F048268: from=<[EMAIL
PROTECTED]>, size=37107, nrcpt=1 (queue active)
Oct 28 15:23:53 vador postfix/qmgr[20995]: 307AB48421: from=<[EMAIL
PROTECTED]>, size=2004, nrcpt=1 (queue active)
Oct 28 15:25:24 vador postfix/qmgr[21984]: 90C67483AE: from=<[EMAIL
PROTECTED]>, size=143536, nrcpt=1 (queue active)
Oct 28 16:11:28 vador postfix/qmgr[9223]: 5A56848492: from=<[EMAIL PROTECTED]>,
size=3253, nrcpt=1 (queue active)
>> >> T 5 10 20 40 80 160 320 640 1280
>> >> 1280+
>> >> TOTAL 3201 2121 937 87 11 21 23 0 0 0
>> >> 1
>> >
>> >Almost all the mail in "incoming" is recent, with ~2000 msgs under 5
>> >minutes old, and ~1000 between 5 and 10 minutes old. What happened
>> >leading up to this?
>>
>> I do not know.
>
>You have to find out, by reading your logs, looking at message rates,
>failed deliveries that took a long time to complete, ...
The logs is 400Mo day... do you have/know programs to do that ?
-rw------- 1 root root 422607285 Oct 28 04:02
/opt/pmx/postfix/var/log/maillog.1
-rw------- 1 root root 331524767 Oct 28 20:36 /opt/pmx/postfix/var/log/maillog
>Is trivial-rewrite performing well? Are you using SQL or LDAP transport
>tables? qmgr itself is only slow when disk I/O is saturated or trivial-
>rewrite is slow.
How can I check trivial-rewrite performance or see if is slow ?
No SQL no LDAP.
transport table
apec.fr smtp:[192.168.100.111]:25
apec.asso.fr smtp:[192.168.100.111]:25
courriercadres.com smtp:[192.168.100.45]:25
>> >> qshape active
>> >>
>> >> T 5 10 20 40 80 160 320 640 1280
>> >> 1280+
>> >> TOTAL 32 0 0 32 0 0 0 0 0 0 >> >> 0
>> >
>> >Exactly 32 active messages, all 10-20 minutes old, what happened leading
>> >up to this?
>> >
>> >What's in the deferred queue? Read your logs, looking for errors, warnings,
>> >throttled transports and nexthops, ...
>> >
>>
>> qshape deferred
>>
>> T 5 10 20 40 80 160 320 640 1280
>> 1280+
>> TOTAL 480 2 0 1 1 13 61 187 36 9
>> 170
>> hotmail.com 152 1 0 1 0 2 38 103 7 0
>> 0
>> hotmail.fr 107 1 0 0 0 8 15 67 16 0
>> 0
>> cji2.net 61 0 0 0 1 3 5 6 1 0
>> 45
>
>Not much.
>
>> transport (same subnet).
>>
>>
>> >Is your disk I/O adequate? ... The information you provide is too skimpy.
>>
>> I have 2x36 u320 15000rpm and 2x73Go u320 15000rpm and 4go memory.
>
>Disk I/O is not disk storage capacity. Is the disk saturated with I/O requests?
The disks are U320 15000rpm with linux md raid 1.
I susppose disk are not satured.
Iif I need I/O I have the I/O, the logs are very bigs.
Linux 2.6.9-78.0.5.ELsmp (vador.apec.fr) 10/28/2008
avg-cpu: %user %nice %sys %iowait %idle
12.66 0.00 2.40 16.39 68.56
Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz
avgqu-sz await svctm %util
sda 0.44 247.84 1.48 112.38 79.39 2043.17 39.69 1021.59
18.64 2.07 18.19 5.53 62.97
sda1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 11.42
0.00 6.78 6.10 0.00
sda2 0.44 247.84 1.48 112.38 79.38 2043.17 39.69 1021.59
18.64 2.07 18.19 5.53 62.97
sdb 0.44 247.82 1.51 112.39 77.87 2043.17 38.93 1021.59
18.62 2.08 18.24 5.48 62.46
sdb1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.70
0.00 8.18 7.05 0.00
sdb2 0.44 247.82 1.51 112.39 77.86 2043.17 38.93 1021.59
18.62 2.08 18.24 5.48 62.46
sdc 0.52 83.76 12.75 42.97 378.59 1023.73 189.29 511.86 25.17
1.02 17.93 2.11 11.74
sdc1 0.52 83.76 12.75 42.97 378.59 1023.73 189.29 511.86 25.17
1.02 17.93 2.11 11.74
sdd 0.52 83.73 12.81 43.06 378.64 1023.73 189.32 511.86 25.10
0.16 2.61 2.09 11.70
sdd1 0.52 83.73 12.81 43.06 378.64 1023.73 189.32 511.86 25.10
0.16 2.61 2.09 11.70
md0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.00
0.00 0.00 0.00 0.00
md2 0.00 0.00 26.60 126.93 757.23 1015.48 378.61 507.74
11.55 0.00 0.00 0.00 0.00
md1 0.00 0.00 3.88 355.38 157.25 2004.38 78.62 1002.19
6.02 0.00 0.00 0.00 0.00
dm-0 0.00 0.00 0.15 2.53 1.69 5.05 0.84 2.53 2.52
0.07 27.68 3.78 1.01
dm-1 0.00 0.00 0.03 0.03 0.26 0.25 0.13 0.12 8.00
0.00 57.90 2.42 0.02
dm-2 0.00 0.00 0.16 0.78 3.55 6.27 1.77 3.13 10.43
0.02 24.06 6.46 0.61
dm-3 0.00 0.00 0.36 4.81 7.76 9.62 3.88 4.81 3.36
0.35 66.88 1.93 1.00
dm-4 0.00 0.00 0.71 132.40 2.43 264.80 1.22 132.40
2.01 1.88 14.13 0.11 1.46
dm-5 0.00 0.00 0.00 0.04 0.02 0.07 0.01 0.04 2.25
0.00 38.79 27.54 0.11
dm-6 0.00 0.00 0.00 0.00 0.02 0.00 0.01 0.00 4.12
0.00 13.29 5.81 0.00
dm-7 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 7.22
0.00 18.66 0.59 0.00
dm-8 0.00 0.00 0.00 0.00 0.02 0.00 0.01 0.00 10.91
0.00 20.35 0.47 0.00
dm-9 0.00 0.00 0.01 0.00 0.02 0.01 0.01 0.00 3.41
0.00 38.66 4.95 0.00
dm-10 0.00 0.00 1.13 56.72 50.30 453.76 25.15 226.88 8.71
0.35 6.13 0.43 2.49
dm-11 0.00 0.00 0.75 120.36 76.79 962.86 38.39 481.43
8.58 1.00 8.25 5.48 66.36
dm-12 0.00 0.00 0.56 37.71 14.39 301.68 7.19 150.84 8.26
0.96 25.01 1.98 7.58
dm-13 0.00 0.00 0.19 10.90 5.23 87.17 2.61 43.58 8.34
1.89 170.48 3.09 3.42
dm-14 0.00 0.00 0.47 4.44 3.82 35.49 1.91 17.75 8.01
0.49 98.97 4.62 2.27
dm-15 0.00 0.00 25.92 111.36 746.91 890.85 373.46 445.42
11.93 5.67 41.33 0.94 12.96
dm-16 0.00 0.00 0.01 0.25 1.24 1.97 0.62 0.99 12.52
0.06 233.57 0.40 0.01
dm-17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.00
0.00 12.65 2.19 0.00
dm-11 = postfix logs
dm-12 = postfix queues
dm-15 = puremessage postgres
dm-10 = postfix/puremessage code, virus/spam sign
>> The information is too skimpy for this we need your help, we are trying for
>> several weeks, lot of configs changes in_flow_delay, default_dest...,
>> smtpd_error_sleep_time = 0s, increase/decrease postfix/puremessage process,
>> install >>local dns... without any success some time this is very bad we are
>> in the fog...
>
>
>You have to make better measurements if you want help. The queue manager
>moves mail from incoming to active very rapidly unless the disk is
>saturated or trivial-rewrite is gummed up, ...
how can I do that with postfix ? how can I know if trivial-rewrite is gummed
up ?
>
>Do review your logs in detail, look at disk I/O, /CPU, memory utilization,
>post your "postconf -n". Run a syscall tracer on the queue manager and
>see where it blocks, ...
>
>--
> Viktor.
>
>Disclaimer: off-list followups get on-list replies or get ignored.
>Please do not ignore the "Reply-To" header.
>
>To unsubscribe from the postfix-users list, visit
>http://www.postfix.org/lists.html or click the link below:
><mailto:[EMAIL PROTECTED]>
>
>If my response solves your problem, the best way to thank me is to not
>send an "it worked, thanks" follow-up. If you must respond, please put
>"It worked, thanks" in the "Subject" so I can delete these quickly.
Best regards.
Francis
