Good morning everyone.
We have a slightly annoying situation on our ibm i, and it's becoming critical, but I don't know how to solve it.
I'll try to explain.
A very powerful user has set up an application on an external server (it's called Ignition), this app allows to collect and manage data from multiple database sources via jdbc drivers. Problem is, this application is querying the iSeries like mad, some queries are launched up to 5 times per second.
All those queries are basically DOSing our IBMi, the sum of QZDASOINIT jobs CPU occupation of this user is always at 30% minimum. Hence the rest of the company is using only the 60/70% of the CPU, causing hangs, delayed jobs, etc.
I've already spoken with an IBM technician and he told me that there isn't an easy solution. He told me about memory pools, but I'm a programmer and have poor experience about system administration.
I'm aware and it's obvious that the problem should be solved from the root, but unfortunately this is not possibile, so I'd like an advice on how should you approach this situation, keeping in consideration that the queries can't be stopped. Nor the dude can be physically harmed or fired.
USRPRF is always the same, jobs are always QZDASOINIT.
We are on a Power8 V7R4, we have only 1 active CPU.
We have a slightly annoying situation on our ibm i, and it's becoming critical, but I don't know how to solve it.
I'll try to explain.
A very powerful user has set up an application on an external server (it's called Ignition), this app allows to collect and manage data from multiple database sources via jdbc drivers. Problem is, this application is querying the iSeries like mad, some queries are launched up to 5 times per second.
All those queries are basically DOSing our IBMi, the sum of QZDASOINIT jobs CPU occupation of this user is always at 30% minimum. Hence the rest of the company is using only the 60/70% of the CPU, causing hangs, delayed jobs, etc.
I've already spoken with an IBM technician and he told me that there isn't an easy solution. He told me about memory pools, but I'm a programmer and have poor experience about system administration.
I'm aware and it's obvious that the problem should be solved from the root, but unfortunately this is not possibile, so I'd like an advice on how should you approach this situation, keeping in consideration that the queries can't be stopped. Nor the dude can be physically harmed or fired.
USRPRF is always the same, jobs are always QZDASOINIT.
We are on a Power8 V7R4, we have only 1 active CPU.
Comment