ASG
IBM
Zystems
Cressida
Icon
Netflexity
 
  MQSeries.net
Search  Search       Tech Exchange      Education      Certifications      Library      Info Center      SupportPacs      LinkedIn  Search  Search                                                                   FAQ  FAQ   Usergroups  Usergroups
 
Register  ::  Log in Log in to check your private messages
 
RSS Feed - WebSphere MQ Support RSS Feed - Message Broker Support

MQSeries.net Forum Index » General IBM MQ Support » [Solved] amqcrata processes

Post new topic  Reply to topic
 [Solved] amqcrata processes « View previous topic :: View next topic » 
Author Message
Balmer
PostPosted: Thu Jul 10, 2003 10:49 am    Post subject: [Solved] amqcrata processes Reply with quote

Novice

Joined: 19 Jun 2003
Posts: 13

hi all,
environment is as follows.
MQ 5.2.1 with latest patch
sunOS 5.8

when see the no of amqcrsta processes using
ps -ef | grep amqcrsta | wc -l, it is giving me 1600. this no keep on increasing and we even not able to logon to the system with mqm user id...

but in /etc/services and /etc/inetd.conf file, we have only three MQ services and three amqcrsta processes configured...


can anyone tell us why the no of amqcrsta processes keep on increasing????

this is on our production environment and solution to this is vere very urgent..

thanks
Balmer
Back to top
View user's profile Send private message
Balmer
PostPosted: Thu Jul 10, 2003 10:52 am    Post subject: Reply with quote

Novice

Joined: 19 Jun 2003
Posts: 13

it is not amcrata process..it is amqcrsta channel listener processes.
Back to top
View user's profile Send private message
tillywern
PostPosted: Thu Jul 10, 2003 1:05 pm    Post subject: Clean them up. Reply with quote

Centurion

Joined: 28 Jan 2003
Posts: 109
Location: Colorado

This is a common problem. With a count as high as 1600 I suspect that you are having a lot of channel activity.

amqcrsta is the channel receiver program. They get started when a channel needs to sent data. The problem is that they don't always go away. As you have seen.

The first thing you can do is look at how your channels are defined. Look at the value of the disconnect interval. High disconnect intervals tell the channels to wait a long time to go away. I have seen this lead to a lot of amqcrstas.

Second is to use the ps-ef command to look at how long the reciver programs have been out there. On one system I was working on they would remain for days. I generated a script that would go out and execute a kill against reciever process that were days ore even hours old.

This might seem like a real hack but I haven't noticed any problems..

You didn't indicate anything about what type of MQ traffic is going in and oit of the machine. Obviously if you were running 1000 MQ clients then the process table might look a bit different than if you have only a ver server to server channels out there.

You might want to take the time to monitor your ipc elements as well. I would wager that with all of these processes out there you are using a lot of shared memory and semaphors.

I would write a script that recorded a copy of the mqseries related processes from the ps command and the ipc elements from te ipcs command. Echo out the date. Run this repeatedly throughout the day via cron or a wile loop and sleep statement. Then you will be able to get an idea of what effect load has on the occurance of processes. If you could start from a point where MQSeries was shut down and all MQ processes and ipc elements were gone you would reall get an understanding of what MQ looks like when you bring it up and then as load increases.

reply back if you need more info on this process.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger MSN Messenger
clindsey
PostPosted: Thu Jul 10, 2003 3:46 pm    Post subject: Reply with quote

Knight

Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx

In addition to what Tillywern advised, I would recommend you set
tcp keepalive for the queue manager. Edit the qm.ini file for the queue manger (/var/mqm/qmgrs/<qmgrname>/qm.ini
Add to TCP & CHANNELS stanzas -
TCP:
KeepAlive=Yes
CHANNELS:
AdoptNewMCA=ALL

You must stop/start the queue manager to pick up the changes.

Before you kill the amqcrsta processes, you can verify that they are orphaned by this method. You may want to automate this.
'echo dis chs(*) JOBNAME | runmqsc qmgrname > list.out'
'ps -ef | grep amqcrsta > ps.out'

list.out will show the PID ( 1st 8 bytes of jobname ). Any entries in ps.out that do not have a match in list.out are orpaned processes.

Hope this helps,
Charlie
Back to top
View user's profile Send private message
mrlinux
PostPosted: Fri Jul 11, 2003 6:57 am    Post subject: Reply with quote

Grand Master

Joined: 14 Feb 2002
Posts: 1261
Location: Detroit,MI USA

Good Tip !!!
_________________
Jeff

IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries
Back to top
View user's profile Send private message Send e-mail
lnm
PostPosted: Tue Jul 22, 2003 5:38 am    Post subject: Reply with quote

Apprentice

Joined: 12 Mar 2002
Posts: 43
Location: Florida

We had some problems that ended up being a rogue application. But IBM was able to provide to us some programs to run that allow you to see what application has how many connections. This is basically the same thing that is being offered in 5.3 as an admin tool.
This was something they provided to us specifically for AIX 5.1 CSD 4 and CSD 6.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic  Reply to topic Page 1 of 1

MQSeries.net Forum Index » General IBM MQ Support » [Solved] amqcrata processes
Jump to:  



You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
Protected by Anti-Spam ACP
 
 


Theme by Dustin Baccetti
Powered by phpBB © 2001, 2002 phpBB Group

Copyright © MQSeries.net. All rights reserved.