On 19.02.2013 18:26, Jim Jagielski wrote: > Hmmm.... I'm not seeing crashes, ...
Concerning the crashes using prefork on Solaris 10. I have a reproduction scenario, but I need to load lots of modules. But then the stacks look very similar to the problem described in http://mail-archives.apache.org/mod_mbox/httpd-dev/200912.mbox/%[email protected]%3E I start the web server, access one static page, gracefully restart and access the same page. I do get the response, but the process after that crashes. The frame #2 0x00045b48 in eor_bucket_cleanup (data=<optimized out>) at eor_bucket.c:37 calls ap_increment_counts() in the scoreboard, and it seems the pfn_ap_logio_get_last_bytes function pointer points to invalid memory. I get a variety of crashes, segfault, illegal instruction etc. It look like again after the restart it uses an old function pointer but the load order of the modules has changed, so the function pointer points to an invalid address. Regards, Rainer
