Quanah Gibson-Mount wrote:
--On Thursday, August 30, 2007 9:19 PM -0700 Quanah Gibson-Mount quanah@zimbra.com wrote:
Thread 15 Crashed: 0 slapd 0x0002d4f0 slap_mods_check + 2292 (modify.c:919)
Hm, in this instance, it is a patched version of modify.c, so the line number doesn't correspond exactly to what is in 2.3.37. :/ On the other hand, this same server crashed similarly with a 2.3.34 build that wasn't patched (but without debugging symbols, so not possible to determine exactly why).
It's pointless to mention a crash in a release you're no longer running, since you obviously can't provide trace info or anything else useful about it any more.
It's inconsiderate to ask for help about a generic release when in fact you've applied private patches to it. You can apply your own patches but that's your decision and your responsibility. Nobody else knows what you've got in your source; we walk into this thread assuming it's one thing but it's something else completely.
(after further investigation on YM reveals the patch is a backport from HEAD...)
Taking it upon yourself to migrate experimental code from CVS HEAD into a production service is risky to begin with, and places you under even greater responsibility. There's a reason we differentiate between the release branch and the development branch. Deploying into a test environment is one thing... You have to stay on top of the CVS commits in case subsequent patches affect the code you're using. (As in this case, where the exact same bug you encountered was fixed in HEAD.)