SV: [PHP4BETA] repeated segfaults at mod_php4.c, anyone else?

> > I recently filed a bug report at: > > http://bugs.php.net/version4/bugs.php?id=2866 > > concerning repeated segfaults in mod_php4.c (php4.0beta 2 & 3). It keeps > > bringing down our entire Apache-server. I still haven't found the exact > > trigger to reproduce it at will, but the server crashes some 20-40 times > > each day... > > > > My questions are: > > 1. Did anybody else experience the same problem? > > 2. Are there archives of old PHP-versions (= 4.0b1) somewhere? > Beta 1 was > > much more stable (for us), and I believe I'm forced to go back > to using it > > until there is a fix/solution for the segfaults. > > 3. Any other suggestions - did I miss anything vital in the > bug-report, can > > I do anything else to help getting a fix for it? > > I bet if you try it on a non-SSL'ed server it won't happen. I have a > feeling that you are compiling PHP against a set of Apache header files > that have not been patched for SSL. > > -Rasmus I've now tried the latest CVS (01/DEC/99) *without* SSL, but httpd keeps on segfaulting. It appears to be (at least partly) duplicating bugs no: 2669, 2848, 2858, 2693 et al., which all report that max_execution_time (in php.ini) is not honored. I experience the same problem. For every segfault there is a log-line reporting: "PHP Fatal error: Maximum execution time of 20 seconds exceeded in [no active file] on line 0" An infinite loop in a PHP-process will: a) keep on running without respect to max_execution_time, b) keep on running even after a browser abort. This code: while (1) { $i ++; } keeps on running, ignoring browser-aborts and timelimit, until it is stopped manually or by a crash. Requesting the above script 10 times in a row leaves 10 infinitely looping processes on the machine. For backtrace of segfault, please see bugreport, URL above. Best regards /Johan Ekenberg