[Resolved] 4.0.16-> 4.0.17 Broke my wordpress blog

L

lugus

Member
#1
#1
Upgraded using cp to 4.0.17. When I went to login to my Wordpress blog it tried to download the file login.php rather than run it. Same with other admin php files. When I went back to 4.0.16 the problem went away.

Any ideas?

PHP 5.3.3 on Centos 5.x
 
Last edited by a moderator:
N

NiteWave

Administrator
#2
#2
looks a serious issue and hard to trace...

what's your wordpress version? I upgraded to 4.017 on a wpmu 2.9.2 site, php 5.3.2, without problem.
 
L

lugus

Member
#3
#3
Version 3 of Wordpress. Is there a chance I need to rebuild php after the upgrade? Have you tinkered with php memory or anything along those lines?
 
N

NiteWave

Administrator
#4
#4
4.0.17 select latest APC 3.1.4 instead of 3.1.3p1 when build PHP in admin console. so if you want to try APC 3.1.4, rebuild PHP in 4.0.17
 
L

lugus

Member
#5
#5
I would if I thought it could help but I do not have APC compiled in my php at all and 4.0.16 works fine without it. Are you saying I have to have it now for 4.0.17?
 
N

NiteWave

Administrator
#6
#6
do you have other opcode cache like eAccelerator, XCache enabled?
if none of them enabled, recommend you have one, absolutely it'll improve the php performance a lot.

opcode cache is built as php extension, you can save/keep your current php binary, just build out apc.so, xcache.so, eaccelerator.so. and enable one(and only one) in php.ini.
 
L

lugus

Member
#7
#7
I do not have any opcode cache compiled at the moment while I experiment. I will when this server goes into full use. I definitely do not want to add it now until I sort out why patch 4.0.17 is breaking the php.
 
L

lugus

Member
#8
#8
Also I have just found if I try and compile php under 4.0.17 I receive:

Fail to generate build script, please try to manually fix the error first.
Failed to create manual install script: /usr/local/lsws/phpbuild/buildphp_manual_run.sh
 
L

lugus

Member
#15
#15
Unfortunately there was an issue with the update that the Litespeed team repaired for me. In the past when I had a similar error using apache it was down to there being insufficient memory allocated to php.
 
N

NiteWave

Administrator
#16
#16
the issue was introduced by 4.0.17 new feature:mod_action support. its PHP Content-Type is "handler/fcgid-script". the new 4.0.17 build ignore "fcgid-script" and fixed the problem.

lugus,
can you give more info about your current problem? you can post here or through PM.
 
L

lugus

Member
#17
#17
Sorry, I did not mean to imply I had an issue now. I should have said, the update originally caused a problem but was cured after work by the Litespeed team.
 
J

Jos

New Member
#18
#18
lugus said:
Upgraded using cp to 4.0.17. When I went to login to my Wordpress blog it tried to download the file login.php rather than run it. Same with other admin php files. When I went back to 4.0.16 the problem went away.

Any ideas?

PHP 5.3.3 on Centos 5.x
Click to expand...
To experience wordpress "WSOD" (white screen of death) when upgrading isn't funny and very timeconsuming. You'll get no clues at all, nothing in the logs. Just a white screen telling nothing.

I did an uprade to lsws 4.0.17 and instantly it killed my blog. Now it is up and running again, but it took a while before finding out what it was which caused the downtime. A bug, hacking or something else?

In my case the lsws upgrade in cooperation with the debian server did corrupt the mysql database by giving it a mix of UTF8 and latin1 encodings which resulted in "funny characters". It wasn't easy to restore the database and reinstall everything. Now lsws 4.0.14 is installed together with php 5.2.13 and it seems to work very well.

Finally after searching the net I learned the path how to get out of the mess. Valuable search terms are: "Turning MySQL data in latin1 to utf8 utf-8" (I cannot link to the actual site due to this is my 1st post in this forum).
 
Last edited:
You must log in or register to reply here.
Top