MQSeries.net :: View topic

sarathmattam · Posted: Sat Jul 04, 2009 2:10 am Post subject: TimeOut Requests are getting Lost

Dear All,

Am facing one weird issue with the Timeout nodes.

I have four message flows deployed in the same execution group (say Default) of the same broker (say DEF_BRK) in Solaris 10 Environment. I am using MB 6.1 and MQ 6.0. All the message flows are doing a consolidation process of the input messages and the consolidated output should be sent to the end system at 11:00 PM every day. It has been working fine for the last one month. But one day, the file got updated in the FTP location but it was 0 kb file where the actual file was still in the mqsitransit folder. I have checked the CURDEPTH of the SYSTEM.BROKER.TIMEOUT.QUEUE and it was 3(where 4 is expected). Now I have cleared all the 4 triggers from SYSTEM.BROKER.TIMEOUT.QUEUE and posted freshly. It worked on the first day and again giving problems on the second day. The curdepth also is reduced to 3 now. Am sure that nobody is deleting the message from SYSTEM.BROKER.TIMEOUT.QUEUE. Also there were no errors reported in /var/adm/messages

Now I have the following doubts

1. What are all the possibilities for such an incident to happen.
2. In case of failure, if I move the output file manually (which is supposed to happen automatically) to the FTP location and delete the file from mqsitransit file, will it affect the trigger in any ways. Because on the days when the trigger failed I used this method to move the file.

Thanks in advance