Hello Cognoise,
We are on a windows server.
I was wondering there is a way to terminate the BiBus (one among many) if exceeds 1.5 GB ? I believe that it is possible through a .bat/.cmd script, but I'm not sure of the solution. I'd really appreciate if some can guide accordingly.
Thanks in advance.
have you investigated already why your bibus service reaches 1.5 GB?
I don't think it's a good idea to force kill bibus all the time.
But im not sure of your situation.
Anyway, i googled and found a script that might help you this one is trying to kill dllhost:
https://www.experts-exchange.com/questions/28681830/Auto-kill-process-when-exceeding-80-memory-consumtion-Windows-2003-server.html (https://www.experts-exchange.com/questions/28681830/Auto-kill-process-when-exceeding-80-memory-consumtion-Windows-2003-server.html)
I´m with sdf on investigating the cause.
That bibus may be sharing a perfectly good running report also getting killed with that.
The cause is one or more rogue reports allocating that much memory.
Hi All,
Thanks for a quick response. I very much appreciate your answers.
There are a lot of users across the organization (may be in thousands). Out of which, a few are report authors spread across the organization in different parts of U.S. Some of the shadow I.T group has developed crappy reports which bring down the server (mainly through local processing). I've set the report timeout limit for both interactive and batch services to 1800 sec (30mins). But there are a lot of reports (not maintained by the BI team) which execute within the aforementioned time frame but does a lot of local processing causing the BiBus to exceed 1.5GB (crippling the server). Having said that, it will be a tedious process to shortlist all the crappy report and tune them (although we don't have enough bandwidth to accommodate such thing).
I understand that killing BiBus would terminate all the threads and PID's associated with it. I'd really appreciate if some can guide me accordingly to mitigate the issue without installing any third party tools.
Thanks
Depending on the amount of servers you have you could use the routing rules functionality to route all rogue reports to a dedicated dispatcher.
That way they will not interfere with normal report operation, and buying you time to clean up where possible.
Perhaps defining some combination of the Cognos package governors on the package(s) they are using could shut them down.
What kind of problems is this creating for you? Are you running out of resources?
Hi All,
We do not have any additional dispatchers. Ours is a standalone installation. Although we have added two new processors, we are running out of resources the moment when BIBus's exceed 1.5 GB.
Thanks.