I wanted to share my ‘ChunkSize’ experience in case it helps someone at some point in the future. I’m running QV11 SR2 and called support with a corrupted QVPR. They helped me restore it, and noticed that my ‘ChunkSize’ was set to 100. They asked me to set the value to 20 and we ended the call (after restarting services). I noticed that night that my QVDS tasks all ran much slower – about 5x slower – than normal. I contacted support the next day and they reset the value back to 100. After a restart everything was back to normal again. The chunk size was the only system change that happened, and I’m certain it caused the performance hit.
ChunkSize can be found in the QV Management Service Config file:
Support said the version prior to mine (I won’t assume they mean QV11 SR1) had some issues with the chunk size being set high, so the solution was to lower it. This is why they told me to set it at 20. After I noticed my issues they asked around and found that my version was reported to work better with a larger setting (100).
Apparently the QVDS and the QVMS communicate via something called a work order (XML based perhaps?). The chunk size is the size of this message.
I figured I would share since I couldn’t find anything online about this setting.
Thanks for sharing, I was hoping someone would actually add some valuable information about this setting.
Workorders have been present in qlikview for a while, I guess is how the QlikView Management Service informs the Distribution Service when it has to execute each task. Workorders are stored in C:\ProgramData\QlikTech\ManagementService\Workorders.
You can actually force a workorder refresh here
But I had no idea about the chunksize setting, what happens if you set it to 200? will tasks work faster? what does that setting control?
Daniel – thanks for the response. Unfortunately, I won’t be able to test setting it to 200. I will add that the setting of 100 was default for my install. Also, when support mentioned they found ‘various errors’ in previous versions of QV when that value was set to high, they meant that tasks would fail with various error messages. I would guess that if you set that number too high you would start to experience the same symptoms. It would be interesting to see if you could increase performance, as you asked, by increasing the value. If you find out more please share.