Skip to:

  1. Skip to Jira Navigation
  2. Skip to Side Navigation
  3. Skip to Main Content

channelsPerSlave>1 configuration with child datasets hit allocator caching issue

Description

In a Thor configuration with multiple channels handling rows with child datasets, the row managers per slave and the allocator caching mechanism conspire to cause the row destructor callback to fail to find it's cached allocator.

This was because each slave channel had it's own cache, but the roxiemem slave row managers were sharing the global one.

Conclusion

None

Activity

Show:

Mark Kelly May 19, 2016 at 7:14 PM

I can recreate if you need

Mark Kelly May 19, 2016 at 7:14 PM

Yes.

Jacob Cobbett-Smith May 19, 2016 at 7:07 PM

Found a couple of workunits like that, but I think the logs have gone.

Was the configuration 1 slave with 8 channels?

Mark Kelly May 19, 2016 at 6:46 PM

Look at WUs from 10.241.40.11:8010
Find two that are ~3.5 sec and ~2:43 min:sec

Jacob Cobbett-Smith May 19, 2016 at 6:39 PM

- do you have the 2 workunits from the performance cluster to look at ?

Fixed
Pinned fields
Click on the next to a field label to start pinning.

Details

Components

Assignee

Reporter

Priority

Compatibility

Point

Fix versions

Pull Request URL

https://github.com/hpcc-systems/HPCC-Platform/pull/8593

Created May 3, 2016 at 4:56 PM
Updated May 19, 2016 at 7:14 PM
Resolved May 4, 2016 at 2:32 PM