| Author |
Message
|
| matuwe |
 Posted: Wed Dec 17, 2008 12:25 am Post subject: semaphore busy errors |
 |
|
Master
Joined: 05 Dec 2007
Posts: 296
|
WebSphere MQ for Linux (x86 platform) 6.0.2.4
MQ threw a AMQ6150 semaphore busy errors along with FDC's, which report
Major Errorcode :- xecL_W_LONG_LOCK_WAIT . My Mq is currently not responding to any commands. Looking at people who had this problem and they suggest increasing the log file, I cannot try these suggestions as MQ is just hanging. Please help
.
Everytime I get this error it is related to my httplistner process. My flowdoesnt use Queue but http input with a msg set for wsdl. I only get like 4000 request per hour, could it mean that increased number of connections to the http input can cause this problem?
Sorry there are two questions here.
1, How can I get MQ to respond?
2. If I am not using Queues, what cauld be causing this problem?
Thanking you in advance |
|
| Back to top |
|
 |
| Vitor |
| Posted: Wed Dec 17, 2008 1:41 am Post subject: Re: semaphore busy errors |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| matuwe wrote: |
WebSphere MQ for Linux (x86 platform) 6.0.2.4
MQ threw a AMQ6150 semaphore busy errors along with FDC's, which report
Major Errorcode :- xecL_W_LONG_LOCK_WAIT . My Mq is currently not responding to any commands. Looking at people who had this problem and they suggest increasing the log file, I cannot try these suggestions as MQ is just hanging. Please help |
Look up the ProbeId in the FDC which should be more specific than the error code. Have you tried some of the more assertive methods of closing the queue manager?
.
| matuwe wrote: |
Everytime I get this error it is related to my httplistner process. My flowdoesnt use Queue but http input with a msg set for wsdl. I only get like 4000 request per hour, could it mean that increased number of connections to the http input can cause this problem?
|
This sounds more like WMB than WMQ. If so, what version?
| matuwe wrote: |
| 1, How can I get MQ to respond? |
I think all you can do is cancel it, change it and reconfigure it
| matuwe wrote: |
| 2. If I am not using Queues, what cauld be causing this problem? |
If you're using WMQ, how are you not using queues? Even if you're using WMB, there are queues underlying it.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 2:20 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
My Probe Id :- XC307070 I have searched for the probe id on google and the solution could be encreasing the log size.
All my flows are not writing or reading to queue, sorry I wasn't specific. Could it be that the httpinput and output write the information to the queues while waiting for response?
How can I cancel MQ so I can reconfigure. It doesn't respond at all to all commands or even write another FDC file to see that it is still alive. Usually they would restart the server, but this time they do not want to do it. My queue manager is simply and standard as created by the crtmqm command. No spacial clustering or stuff. Please help, I just need to get MQ responding |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Dec 17, 2008 2:33 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| matuwe wrote: |
| All my flows are not writing or reading to queue, sorry I wasn't specific. Could it be that the httpinput and output write the information to the queues while waiting for response? |
Well something's causing the queue manager to access the logs.
| matuwe wrote: |
| How can I cancel MQ so I can reconfigure. |
There is a procedure in the System Admin manual for manually cancelling a queue manager. Be sure you've exhausted all other options.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Wed Dec 17, 2008 4:03 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
Yes, the bipHTTPListener uses MQ to communicate with the HTTP nodes.
You need to end your broker, completely.
Then, and only then, will you have a reasonable chance of stopping your queue manager in a semi-safe way.
Your best bet at this point, really, is to disable any automatic startup options you have configured or scripted for broker and MQ, and then reboot the machine.
Then you can look at repairing your queue manager. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 4:24 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
Owwww that explains it, I am using Httpinput and Fileinput, for some reason fileinput is still working even though MQ has frozen.
I know rebooting will fix my problem, but that is not an option. They have restarted the pc 3 times in the past and do not want to do it any more.  what can I say it's the client.
I tried to run the netstat -an | grep 7080 command to check if I still have connections to the httpnput, my results
tcp 0 0 :::7080 :::* LISTEN
tcp 1 0 ::ffff:185.234.201.17:7080 ::ffff:185.234.201.11:26139 CLOSE_WAIT
tcp 1 0 ::ffff:185.234.201.17:7080 ::ffff:185.234.201.11:36890 CLOSE_WAIT
tcp 1 0 ::ffff:185.234.201.17:7080 ::ffff:185.234.201.11:26137 CLOSE_WAIT
tcp 1 0 ::ffff:185.234.201.17:7080 ::ffff:185.234.201.11:26136 CLOSE_WAIT
There are lots of CLOSE_WAIT, what does it mean? I tried to kill the execution group that has the httpint and I stopped the client application. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 4:30 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
| should I perhaps kill my biphttplistener process? |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Dec 17, 2008 4:31 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| matuwe wrote: |
| I tried to kill the execution group that has the httpint and I stopped the client application. |
Shut down the broker, as previously suggested.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 4:44 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
I cannot? Unless if I do the mqsistop -i . But I am Nervous that this problem is on my production enviroment, so currently I have one interface down (httpinput) but 12 others that are using Fileinput are working. If I force broker to stop, I know I will not be able to bring it up until MQ is fixed, I can only try it after hours. .
I will try it later when all the users are gone home. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 5:11 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
After restarting broker na MQ (Once my problem is resolved) How can I make sure I do not have this problem happening again.
Will encreasing the MQ log help. Or do I have to create another execution group and deploy another copy of my httpinput flow to try and distribute the load? |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 7:29 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
I finally have a chance to stop broker. The mqsistop -i Broker name doesn't stop the broker. I have killed -9 the process id's and now dspmq still doesn't respond.  please help |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Dec 17, 2008 7:49 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| matuwe wrote: |
| please help |
Obtain permission to reboot the pc.
If permission is denied, arrange for the mains cable to be pulled out of the socket in a freak, million-to-one accident.
It always works for me. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 7:56 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
 i like that. But the problem is, this is not the first time it happened that is why they are not happy with reboot. Last time I was lucky enough the whole building power went down. But I do not think I will have that kind of luck again.
Now MQ is not writing FDC or errors.. It is just silent. Is rebooting my only option? I might as well get into trouble, than have the system not working tomorrow morning... |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Dec 17, 2008 8:03 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| matuwe wrote: |
| Last time I was lucky enough the whole building power went down. |
I like your style! Covering your tracks with a more generalised problem; nice. 
| matuwe wrote: |
| Now MQ is not writing FDC or errors.. It is just silent. Is rebooting my only option? |
It sounds like it's dead. 
| matuwe wrote: |
| I might as well get into trouble, than have the system not working tomorrow morning... |
Well what's going to cause you the most trouble? An unauthorised reboot that gets all the services working, or a system that still partially crippled?
If the clients are complaining at having to continually reboot their PC, advise them to run a real OS not Windoze!
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| matuwe |
| Posted: Wed Dec 17, 2008 8:26 am Post subject: |
|
|
Master
Joined: 05 Dec 2007
Posts: 296
|
Thanks, some how you gave me the courage to tell the client about the problem. They have accepted to restart the server. So hopefully, MQ comes up peacefully after the reboot.
 thanks a mill.... will check if I still have my job 2morrow morning  |
|
| Back to top |
|
|
|
|
