local delivery slow or stops

[sanjayacharya]

I get the following errors

omshowlog -s ld

Local Delivery(Local Delivery) 11.08.07 17:03:05
[OM 24520] The PDL contains no members, hence expansion will not occur.
Last Msg Id: 200711082341.lA8Nfx80003106(a)mail.itsupport247.net
Last Msg DirectRef: 002167226176ad33

ERROR Local Delivery(Local Delivery) 11.10.07 06:28:26
[OM 24520] The PDL contains no members, hence expansion will not occur.
Last Msg Id: 10813.13661194704780.mx1.zenithinfotech.com
Last Msg DirectRef: 0024892ba399bb8a

Please help.

Sanjay

[sanjayacharya]

Some more logs

Current errno value: 3
Last Msg Id: 002001c8246e(036)5a297cd0(036)b603010a(a)zenithnoc.com
Last Msg DirectRef: 000d639eb24392c3
<- sfl_GetNewBlock
-> sfl_GetNewBlock
<- sfl_GetNewBlock
-> sfl_GetNewBlock
<- sfl_GetNewBlock
-> sfl_GetNewBlock
<- sfl_GetNewBlock
-> ct_AdjustCounts
<- ct_AdjustCounts
<- /build/11.0.4/src/lib/ombase/sfl/sfl_monit.c:160[116,380]
<- /build/11.0.4/src/lib/ct/ct_atct.c:959[116,380]
<- /build/11.0.4/src/lib/ct/ct_atct.c:975[116,380]
<- /build/11.0.4/src/lib/ct/ct_atct.c:1008[116,380]
<- /build/11.0.4/src/lib/ct/ct_atct.c:454[116,380]
<- /build/11.0.4/src/bin/ld/ld_2user.c:648[3,7743]
<- /build/11.0.4/src/bin/ld/ld_main.c:1280[3,7743]

omstat -q local shows 300 plus mails

[sanjayacharya]

Is there a way to flush the local queue

[Valerion]

There was an earlier issue on the forum where someone had a PDL with no members causing Local Delivery to abort. Check this, and restart Local Delivery to empty the queue.

[sanjayacharya]

I assume PDL is a distribution list, we have over 50 DLs. is that a problem? Also how can I flush the local delivery . I run omscan -Aavfx once every week.

[Valerion]

Number of DL's isn't the problem, Scalix scales very well in that regard. Just go through each and verify that they all have at least one member (even if it is sxadmin).

To delete the queue contents the easiest is omqdump (password the same as omcontain discussed elsewhere in the forum). Be careful, it is a pwerful tool and can cause some damage if you don't understand it.

What I normally do is to move the contents to a queue like POISON that doesn't get processed, then move them back when I want them delivered.

[sanjayacharya]

every day early morning local mails queue is not empty it ranges between 50 to 1000 and some times more.

Here is the output of tail -n 25 /var/opt/scalix/s/?/logs/fata.
--------------------

Pid of logging process: 24944
Last Msg Id: 32162166.10771195069357906.JavaMail.root(a)mx1.zenithinfotech.com
Last Msg DirectRef: 00017ba4f2459357

ERROR Indexer (Indexer ) Mon Nov 19 00:43:51 2007
[OM 16082] Directory is unavailable at this time. Please retry later.
Pid of logging process: 32557
Current errno value: 4

ERROR Omscan Server (Omscan Server ) Mon Nov 19 02:36:08 2007
[OM 28875] Attempt to read a block which does not exist from a blocked item.
User Name: Masoor Khan / mx1/CN=Masoor Khan
Pid of logging process: 32716

ERROR Browser (Service 14 ) Mon Nov 19 05:56:29 2007
[OM 0]
Pid of logging process: 8388
Last Msg Id: 27500239.10841195071042186.JavaMail.root(a)mx1.zenithinfotech.com
Last Msg DirectRef: 000859432f67202b
-----------------
MY server has two 500GB SATA HDD in RAID 1, is drive writing speed is a issue here.

Sanjay

[sanjayacharya]

Also note that when I run mailq there are only 5 mails in the queue, when i run omstat -q local it shows 900 mails.
Sanjay

[mikethebike]

omsearch -e s=*/entry-typee=4/dl-members= -m s/ou1

any that are returned, are empty, delete the pdl

omdelpdl -l pdlname

omscan will not empty the local delivery queue. If you think its slow increase the number of aux processes
omstat -s -p
omsetsvc -x local <number of aux processes>

see the man page for more details, also detailed in teh technical guide.

having 5 mails in teh smtp queue does not mean there is a problem, you would need to look in the logs to see if they have been delayed (if its sendmail you can look at the "q" file to see the routing history of that message).

Mick

[sanjayacharya]

I have checked all PDLs and none of them are empty, aparently it looks like I have I/O issue which causes local delivery to be slow. I have a 4Gb server with 400Gb X2 Sata Drives mirrored.there are about 800 Users and server processes atleast a 10000 mails a day.

Any suggestion on improving the I/O on a Redhat .

[mikethebike]

If you have a busy system, make sure you have seperate mount points and physical disks for the busier filesystems.
~temp
~data
~user*

Mick