| Author |
Message
|
| siliconfish |
 Posted: Mon Feb 17, 2003 12:01 am Post subject: URGENT Production issue : runmqlsr onWindows NT |
 |
|
Master
Joined: 12 Aug 2002
Posts: 203
Location: USA
|
This Scenario is in Windows NT.
I have a Q Mgr TEST_QMGR on which runmqlsr is running on 1414.
Due to some reason runmqlsr stopped and when we started the runmqlsr on the port 1414 again, we are getting error : 1414 is in use.
we issued "netstat -a" command and confirmed 1414 is in use.
I know that when we recycle the Queue Manager and issue runmqlsr again this will go of. This, I have tested on my test box.
But Stopping the Queue Manager is not the option we want and also there is nothing mentioned in the AMQERR logs.
But I want to find out if there is anyway to freeup the port in NT which is in use by an abnormally ended or corrupeted process which didn't close the port.
Can somebbody also tell me how to free up the port in use inSolaris also.
Thanks in advance. 
_________________
siliconfish |
|
| Back to top |
|
 |
| LuisFer |
| Posted: Mon Feb 17, 2003 12:11 am Post subject: Re: URGENT Production issue : runmqlsr onWindows NT |
|
|
Partisan
Joined: 17 Aug 2002
Posts: 302
|
| I think that there is a PTF for this problem. |
|
| Back to top |
|
|
| siliconfish |
| Posted: Mon Feb 17, 2003 12:21 am Post subject: |
|
|
Master
Joined: 12 Aug 2002
Posts: 203
Location: USA
|
What is PTF?
_________________
siliconfish |
|
| Back to top |
|
|
| dgolding |
| Posted: Mon Feb 17, 2003 12:24 am Post subject: |
|
|
Yatiri
Joined: 16 May 2001
Posts: 668
Location: Switzerland
|
PTF=Program technical Fix or is it Temporary fix? Anyway, it is a patch from IBM.
Here's some more jargon, CSDs contain all all the known fixes to the product - so it's a good idea to get the latest CSD for version of NT and apply it. You can get it off the IBM web-site.
HTH |
|
| Back to top |
|
|
| siliconfish |
| Posted: Mon Feb 17, 2003 2:14 am Post subject: |
|
|
Master
Joined: 12 Aug 2002
Posts: 203
Location: USA
|
We have already applied CSD05.
_________________
siliconfish |
|
| Back to top |
|
|
| vmcgloin |
| Posted: Mon Feb 17, 2003 3:45 am Post subject: |
|
|
Knight
Joined: 04 Apr 2002
Posts: 560
Location: Scotland
|
What does netstat -a show for the port when the problem occurs?
If you are seeing the same on Solaris then it could be normal tcp/ip behaviour.
Cheers,
Vicky |
|
| Back to top |
|
|
| pgorak |
| Posted: Mon Feb 17, 2003 5:26 am Post subject: |
|
|
Disciple
Joined: 15 Jul 2002
Posts: 158
Location: Cracow, Poland
|
| Quote: |
| Due to some reason runmqlsr stopped |
For what reason, then? This definitely should not end abnormally. If it does (e.g. someone kills the process) the socket will stay in FIN_WAIT state for some time and after that the port will be released.
Piotr |
|
| Back to top |
|
|
| bbeardsley |
| Posted: Mon Feb 17, 2003 6:27 am Post subject: |
|
|
Acolyte
Joined: 17 Dec 2001
Posts: 52
Location: Dallas, TX, USA
|
| Do you have another queue manager on the same machine? It could have a listener automatically starting up. |
|
| Back to top |
|
|
| siliconfish |
| Posted: Mon Feb 17, 2003 12:06 pm Post subject: |
|
|
Master
Joined: 12 Aug 2002
Posts: 203
Location: USA
|
We don't know the reason for abnormal ending of runmqlsr. This is just my guess as the port is in use, but I haven't got any info from the AMQERR logs, I have checked thoeoughly.
We don't have another Queue Manager on that NT machine.
Anyway we rebooted the machine and now everything is fine.
But Still I don't have the answer to the question : Is there any utility to free up the ports which are in use.
Regarding Solaris, I was just curious to know if there is any utility to do the same. We don't have Queue Managers on solaris.
I also have one more Question if I start multiple listeners at different ports for the same Queue Manager, and define channels from other Queue Managers to this Queue Manager using different ports on the CONNNAME param of all the other Queue Managers, is there any advantage?
Thanks for all the people who have replied.
_________________
siliconfish |
|
| Back to top |
|
|
| clindsey |
| Posted: Mon Feb 17, 2003 1:16 pm Post subject: |
|
|
Knight
Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx
|
Siliconfish,
I would recommend you contact IBM support on this one.
There is a fix in the 5.3 code to reuse the ports when they are left in TIME_WAIT state. The apar number is IY33049. Support can confirm that you have the same problem and possibly get an efix for your level of code.
Also, there may be *FDC files in /var/mqm/errors. If yes, this may help to identify why runmqlsr is terminating.
Charlie |
|
| Back to top |
|
|
| LuisFer |
| Posted: Mon Feb 17, 2003 8:59 pm Post subject: |
|
|
Partisan
Joined: 17 Aug 2002
Posts: 302
|
| There is not problem to start a new LSTR on other port and change the CONNAME to this one. |
|
| Back to top |
|
|
| pgorak |
| Posted: Tue Feb 18, 2003 1:51 am Post subject: |
|
|
Disciple
Joined: 15 Jul 2002
Posts: 158
Location: Cracow, Poland
|
| Quote: |
| I also have one more Question if I start multiple listeners at different ports for the same Queue Manager, and define channels from other Queue Managers to this Queue Manager using different ports on the CONNNAME param of all the other Queue Managers, is there any advantage? |
No. First of all, it is not recommended. Secondly, I did a test some time ago to learn if there are any performance gains when one uses two or more listeners connected to the same queue manager and it turned out there aren't any (during the test a large number of messages was sent to/from the queue manager and the applications involved did little processing save MQGET and MQPUT operations)
Piotr |
|
| Back to top |
|
|
| siliconfish |
| Posted: Tue Feb 18, 2003 2:16 am Post subject: |
|
|
Master
Joined: 12 Aug 2002
Posts: 203
Location: USA
|
Thanks All.
_________________
siliconfish |
|
| Back to top |
|
|
|
|
