Hi Stefan,
yes this seems to correct BUT i'm not sure i the test was long enough
where i hadn't a segfault. I'll rerun a test with that version. To
verify this was no just a "fault".
Greets,
Stefan
Am 05.03.2017 um 14:34 schrieb Stefan Eissing:
> Thanks. I try to summarize:
>
> - we did see this
Thanks. I try to summarize:
- we did see this rarely with unpatched v1.9.2 and Yann's mpm patch v??
- we still see this rarely with the patch adding mutex to the main conn
allocator (so this seems not to change anything)
- we did not see this with v1.9.2 and Yann's patch on ptrans version ??
Is