Re: release v1.9.2

Am 09.03.2017 um 18:09 schrieb Stefan Eissing: > Thanks for running our patches with many changes! Not a mistake to have it > running fine for four days. ;-) > > Hmm, so we hunt a rare beast. After much effort from all sides we made it > rare. So, nothing to be ashamed of. Need to keep

Re: release v1.9.2

Thanks for running our patches with many changes! Not a mistake to have it running fine for four days. ;-) Hmm, so we hunt a rare beast. After much effort from all sides we made it rare. So, nothing to be ashamed of. Need to keep looking... > Am 09.03.2017 um 17:32 schrieb Stefan Priebe -

Re: release v1.9.2

Hi Stefan, while doing longer testing i can say that also the version which was working for 4 days segfaults. So it was never solved ;-( Sorry about that. Testing was just too short. Greets, Stefan Am 05.03.2017 um 14:49 schrieb Stefan Priebe - Profihost AG: > Hi Stefan, > > yes this seems to

Re: release v1.9.2

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

Re: release v1.9.2

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 ??

Re: release v1.9.2

Hi Stefan, current trace - not sure whether this is http2 related at all: Core was generated by `/usr/local/apache/bin/httpd -DFOREGROUND'. Program terminated with signal SIGSEGV, Segmentation fault. #0 allocator_free (node=0x0, allocator=0x7fbf4c05a410) at memory/unix/apr_pools.c:381 #0

Re: release v1.9.2

Hi Stefan, live. Sorry for the late reply. Stefan Am 28.02.2017 um 11:49 schrieb Stefan Eissing: > Meh, my mistake. Sorry about that. Did not cleanup properly. Dare you with an > improved version: > > > > > >> Am 28.02.2017 um 08:41 schrieb Stefan Priebe - Profihost AG >>

Re: release v1.9.2

Meh, my mistake. Sorry about that. Did not cleanup properly. Dare you with an improved version: h2_192_alloc_mutex_v2.diff Description: Binary data > Am 28.02.2017 um 08:41 schrieb Stefan Priebe - Profihost AG > : > > That one breaks everyting. Multiple crashes per

Re: release v1.9.2

That one breaks everyting. Multiple crashes per second. [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by `/usr/local/apache/bin/httpd -DFOREGROUND'. Program terminated with signal SIGABRT, Aborted. #0

Re: release v1.9.2

Damnit. Can you try the attached patch on top of mod_http2 v1.9.2? This one sets a mutex on the main connection allocator if missing and is pretty close to the version we ran successfully with on your site for days. Thanks again! -Stefan h2_192_alloc_mutex.diff Description: Binary data >

Re: release v1.9.2

Hi Stefan, two new crashes here. Reading symbols from /usr/local/apache/bin/httpd...Reading symbols from /usr/lib/debug//usr/local/apache2/bin/httpd...done. done. (gdb) (gdb) quit Reading symbols from /usr/local/apache/bin/httpd...Reading symbols from

Re: release v1.9.2

Hi Stefan, currently everything fine. No segfaults. Stefan Am 25.02.2017 um 20:40 schrieb Stefan Priebe - Profihost AG: > Hi Stefan, > Am 25.02.2017 um 13:51 schrieb Stefan Eissing: >> Stefan, >> >> whenever you have time, please deploy >> https://github.com/icing/mod_h2/releases/tag/v1.9.2 >>