I have a significant issue at the moment with a number of Qlikview tasks that are hanging in the middle of a reload and not completing. Since 1:00am this morning we should have had about 100 tasks reloaded (some of the tasks reload every hour, while some are just one offs) and because some of the tasks are hanging it has resulted in about 70 of the tasks failing to complete. Nothing has changed from a Qlikview point of view (I'm not sure whether other issues could affect this) and these jobs have been running for years now with no problems.
- The tasks that seem to be the problem do an incremental load (although not all incremental loads are failing)
- We use QV 10 SR3 and have a cluster environment
- I don't think it is an issue with Publisher as I've manually run through the development application and same thing exists
- Publisher server has been restarted and have had no effect at all.
If anyone has ever encountered a situation like this please let me know!
Not that I'm aware of, and some of the jobs that have failed bear no relation to each other (i.e. they don't go near the same data set). It seems to be an issue on files that are doing incremental loads and concatenating onto 4M+ rows.
I had something similar recently. Publisher usually executes reload in 30 min, and distributes (copies to another machine) in two or three additional minutes. In one case last week, the distribution portion took more than 30 min, although the reload took the expected time.
I don't know what's the reason. Worked fine before and after. Maybe some network issue (?)
I've managed to identify the reason why the jobs were hanging:
There was an issue on the server (this is something that I'm not clear of yet but essentially is the root of everything) and a couple of tasks that were running basically corrupted the qvds that they were being stored to (looks like they were only partially saved as the usual file size 2GB but it was just 650MB this time). As these qvds were used elsewhere it caused an adverse effect on other jobs which caused the rebuilds to hang
Another thing that has happened since the 1am server problem on Saturday morning is that a qvd that usually gets overwritten (again around 2GB) is failing on the store. I removed the qvd that it was overwriting and it worked fine but seems to fail if the qvd already exists