All Sites suddenly generating "Premature end of script headers" errors

UpAllNight

Verified User
Joined
May 23, 2005
Messages
45
Sorry for the double post. I think I originally posted this under the wrong topic:

Have a number of sites on a server that are using a script based software product called Aestiva. Each of these sites uses a "ScriptAlias" entry in the site's Custom Httpd Configuration like:

ScriptAlias /h20/ /home/username/domains/theusername.com/h20/

These sites have been running on this DA based server ->for years without issue<- but ->as of this morning, 1/3/2009<- they have all stopped working. Trying to access pages that are utilizing this software for every site returns:

"Premature end of script headers: /home/username/domains/theusername.com/h20/start"

I verified the files are stil there, permissions and ownership are correct.

I also tried putting a straight .html file in the alias directory but it still generates a "Premature end of script headers" error in the domain's error log and a "500 internal server error" web page.

(The above has always worked before and was a quick check for me to make sure the ScriptAlias directory was being recognized which now leads me to believe that somewhere the ScriptAlias is not getting picked up by Apache??)

I'm just guessing, but it seems to have been something that changed universally on the server today that is preventing the Custom Httpd Configurations from being read???

I've also tried rebooting the server and manually updating DirectAdmin. All with no change.

Any ideas on what to try??

Thanks,
David
 
I am not familiar with Aestiva but a 500 error is the result of a cgi abort so I think ScriptAlias is not the problem and you need to dig into the apache logs to see what info there is on the failure. The 'premature * and 'internal server' messages are client side messages and are not supposed to tell you what went wrong. There are a ton of reasons why apache will abort a cgi and with a binary cgi like you have it is hard to guess what sort of limit was breached. Remember suexec could also be a player in this this so be sure to check the suexec log too.

Knowing absolutely nothing about your setup except that it has worked for a long time, and assuming you have not made any changes link tinkering with .htaccess or the like, and knowing that the problem survived a reboot, I'd be looking at what is on the disk, maybe log rotation is broken and you have a huge file? Or something the like.
 
Last edited:
Hi gaw,

Smtalk found the problem! The server is running Apache 1.3 and the suexec_log had grown too large for that verions of Apache. He corrected the problem it is working again.

Appreciate your reply and suggestions.
 
Back
Top