Hide Forgot
Created attachment 496779 [details] gdb backtrace 1 Description of problem: I encounter random segfault and 100%CPU on mod_proxy/mod_security/worker configuration, one per hour approximatively. I can get 2 different backtrace, maybe there is two seperate bugs. Thx for your help. Hervé.
Created attachment 496780 [details] gdb backtrace 2
Red Hat Enterprise Linux Server release 5.3 (Tikanga) Linux SI-RP2 2.6.18-128.1.1.el5 #1 SMP Mon Jan 26 13:59:00 EST 2009 i686 i686 i386 GNU/Linux httpd-2.2.3-45.el5 apr-util-1.2.7-11.el5_5.2 apr-1.2.7-11.el5_5.3 mod_security-2.5.12-1.el5
Thanks for the report. The first segfault is triggered within mod_security, and it is quite likely to be related to mod_security. mod_security is not distributed with RHEL. #1 0x008854b0 in ap_pass_brigade (next=0x1, bb=0xa6b7eb50) #2 0x0088c165 in ap_byterange_filter (f=0xa24e340, bb=0xa6b7eb50) #3 0x008854b0 in ap_pass_brigade (next=0x1, bb=0xa6b7eb50) #4 0x00ad0554 in pdfp_output_filter (f=0xa24f220, bb_in=0xa6b7eb50) at pdf_protect.c:360 #5 0x008854b0 in ap_pass_brigade (next=0x1, bb=0xa6b7eb50) #6 0x00af2aae in send_of_brigade (msr=0xa24e9e8, f=0x0) at apache2_io.c:408 The second segfault is in the proxy code. The packages here: http://people.redhat.com/jorton/Tikanga-httpd/ have some possibly-related fixes to the proxy; could you try those?
I just installed your package, we just have to wait for a few hour :) for mod_security, it is the EPEL package, is there a separate bugtracker for it ? Thx for your help.
(In reply to comment #4) > for mod_security, it is the EPEL package, is there a separate bugtracker for it > ? Use the "Fedora EPEL" product here, yes.
I encounter no segfault since i have updated, it seems to correct my issue. Thanks !
Thanks for testing it out. I am going to dupe this against the proxy bug on the presumption you are seeing the same issue. *** This bug has been marked as a duplicate of bug 552303 ***