Atomicorp

Ha!

I see.

I wonder why they still list the flag in their compile configuration list after they took the code out.

Confusing

It used to be activated by default before and no longer is a default that's all.

Well it makes me a little bit nervous to add a feature that the vendor of eaccelerator deactivated.
There is definitely a reason for that. It used to be in there, but it's not anymore.
I don't think this should be compiled in a stable server release.
Maybe you can try SMF 2.0 Release Candidate 3
or
disable eaccelerator only for that domain SMF is hosted on.

You have a point Bruce but I also think that if we only had defaults on, then we would be missing tons of stuff and also, I am still trying to understand why it is listed in their compile options if it had been taken out. (I know they switched it from default to option as the page says) but according to Scott, the code isn't even in the distribution at all which doesn't add up.

Anyway, not the end of the world and perhaps they just haven't updated their page since the other flag mentioned in this thread is not even listed there.

With this new eaccelerator i have many problems on my typo3 sites, the eaccelerator crashes.. !!

Scott, can you tell me, what the difference between:

php-eaccelerator-5.2.13_0.9.6-2.3.el5.art.x86_64.rpm Mar 11 2010 10:18:02

and

php-eaccelerator-5.2.13_0.9.6-3.el5.art.x86_64.rpm Mar 11 2010 17:27:24


last version (3.el5.) seems to be run fine. No Probs with TYPO3.

you can find it here:
http://www.atomicorp.com/channels/atomi ... rator.html
also a patch is included to stop the segfaults discussed here: http://atomicorp.com/forums/viewtopic.php?f=12&t=3926

It seems the --with-eaccelerator-doc-comment-inclusion flag is causing the segfaults reported in http://www.atomicorp.com/forum/viewtopic.php?f=2&t=3953 I just spent 2.5 hours on IM with Scott debugging this.

-4 will be up shortly. Everyone other than breun is banned from making suggestions about this package.

Actually, if I may make a sugg[ARGH!][STOPPIT][OUCH!]....

On a serious note, thanks to Breun and Scott for their efforts on this.

Faris.

I tried 1:5.2.13_0.9.6-4.el5.art and still had Segmentation faults, though fewer.

[Sat Mar 13 08:53:32 2010] [notice] caught SIGTERM, shutting down
[Sat Mar 13 08:53:33 2010] [notice] Apache/2.2.3 (CentOS) configured -- resuming normal operations
[Sat Mar 13 09:48:24 2010] [notice] child pid 5003 exit signal Segmentation fault (11)
[Sat Mar 13 09:48:24 2010] [notice] child pid 5724 exit signal Segmentation fault (11)
[Sat Mar 13 09:48:29 2010] [notice] child pid 4991 exit signal Segmentation fault (11)
[Sat Mar 13 09:48:29 2010] [notice] child pid 5013 exit signal Segmentation fault (11)
[Sat Mar 13 09:48:45 2010] [notice] child pid 4998 exit signal Segmentation fault (11)
[Sat Mar 13 09:49:10 2010] [notice] child pid 5727 exit signal Segmentation fault (11)
[Sat Mar 13 09:49:16 2010] [notice] child pid 4989 exit signal Segmentation fault (11)
[Sat Mar 13 09:49:21 2010] [notice] child pid 5009 exit signal Segmentation fault (11)

You might need to clean out /var/cache/php-eaccelerator after upgrading.

cleaning out /var/cache/php-eaccelerator after upgrading fixed my problems, thanks!

Hi Scott,

and thanks to you and breun taking care of the segfaults the people had.
The verison -3 worked for me because of the eaccelerator patch I send to you.
In version -4 the patch seems not to be compiled in and the segfaults start again.
Did you take the patch out? Is it possible to compile a version in a way where
the conclusion you and breun made are integrated and my patch as well?
Thanks

Hm, -3 was segfaulting for us, -4 isn't. Did you clean out /var/cache/php-eaccelerator after upgrading to -4?