zulfi123786 wrote:The log shows that pid's are created and then the log doesn't get updated so you can take the player pid's and go to the node where they are running and run ps command to see if all those pid's keep running or did the player pid's finish after execution and then check if the section leader finish too.
Its basically a trial and error approach to narrow down to the root cause as to what is causing the issue and even if its found the solution of which might end with PMR
Thanks Zulfi...
I checked , all the PID's are active. All player pid's and event he section leader pid were active. No more info getting on this.
Only a section of jobs around 35 sequence jobs are giving trouble. Rest all jobs in the project are running fine.
Well !!! In digging more to find root cause for this, interstingly, it was found that, we are unable to view the data through designer of any of the sequential files. While we try to do.. after a while [Not Responding] message is appearing and the desinger is closing.
I tried to view data of different files related to different jobs, but they are OK. We checked the files permissions , folder permissions, etc.. all are OK.
Why specifically only these jobs are giving trouble ?
Is it required to clean the log of the jobs ? Might be almost from past 8 months, we are either puging/deleting the logs of these jobs.
we checked the TMP, TEMP folders , those are having enought 25+ GB space.
Well. The final step which I took is to replace the link with new one. It worked. Does this indicates the stage/link got corrupted ?
Well another intresting thing which I noticed is that, in general when we keep a sequentail file stage and either if we are reading from or writing into, the link parition symbol will be "specific parition marker" . But the jobs which got hang were having "Auto paritition" blue square box.
When I replaced the links with new ones.. the jobs are running fine withough any issues.
How come the parition is changed only for these jobs. I checked rest of the jobs in the project where sequential file stage were present, none of them were changed.
rsripathy wrote:Well another intresting thing which I noticed is that, in general when we keep a sequentail file stage and either if we are reading from or writing into, the link parition symbol will be "specific parition marker" . But the jobs which got hang were having "Auto paritition" blue square box.
This happens when you read the file parallely on multiple readers per node or read from multiple nodes, which was never mentioned in your posts. These are not enabled automatically and the designer must have explicitly set them for a purpose. Are your files fixed width ?
rsripathy wrote:When in Designer , this property can check/uncheck ?
I assume the when of above is where. Well just go to the properties page and you will find the option at the bottom go ahead and remove the property. And I still hold on to saying, the property is not a default and should have been explicitly set.