So this thread has been going on for two and a half years now; I was going to start a new one, but I'll just post here.

I get a lot of 500 errors with MT4, especially when I'm publishing, but it can happen at any time. In every case, I can just reload and it works, but I'd love it if things worked the first time, at least most of the time. My error log says 'premature end of script headers at /home/cgi-bin/mt/mt.cgi' -- same as a lot of other people, apparently.

I have MT installed on a server with IX Web Hosting and one on BlueHost -- same issue with both. BlueHost uses cPanel X, but I think IX has something different; at least it looks different to me. I'm wondering if it has something to do with my upgrade from MT3.whatever as opposed to doing a clean install. I didn't ever have this problem before upgrading to MT4. I use MySQL databases.

Any issue particular to cPanel/BlueHost I'm not really concerned with, because they're losing my business -- I'll have the rest of my sites out of there within the next few months. Hate those guys.

It seems like this was a common problem in MT3, but what about MT4?

We are getting the same here, especially when trying to upgrade. This is a snippet from our hosting provider:

"I'm running into another problem with MT 4.x, where even 64MB is not enough... I have seen MT 4.x become so bloated that there are instances to where each script execution wanted between 64 to 128MB per execution - which is completely nuts! I'm pretty sure that MT is going to get a lot of pushback from hosts once their servers start crashing due to OOM...

All in all, the memory limit is there to prevent buggy or bloated scripts from taking your web server down... 32MB is a reasonable (per execution) PHP limit for any shared Community Server, because there is no way the server can discern if your PHP script is only executed by you, or if it is public and in turn cause harm..."

Getting the same problem here, 500 Internal Server Error message:
"The server encountered an unexpected condition which prevented it from fulfilling the request.
The script had an error or it did not produce any output. If there was an error, you should be able to see it in the error log."

I installed MT 4.01 (fresh, not an upgrade) about a month or so ago, rebuilt templates and started populating the site. I've done about 50-60 entries without an problem. However yesterday I started getting the error message above whenever I do a "Publish Site" (which I do fairly often as I use my categories as my menu so I like it to be updated consistently throughout the site).

The problem seems to be intermittent, it'll do it 3x in a row, then it will work. It won't work, I'll delete an entry, then it will work. I'll put in a new entry, it will work. Then I'll do another, it won't. Try again, it will. There seems to be no rhyme or reason to it. Doesn't seem to matter the content of the entry.

I understand this could be a memory usage issue with my web host? I'm using Powweb, not sure if anyone is familiar with them. If this DOES seem to be the problem, can anyone recommend a host that is MT friendly? Is there any kind of temporary work-around, besides not doing a site rebuild?

I've put so much time into this, I'd hate to not have it work right.

I have experienced the same issue with MT 4.1. After lengthy investigations and using MT Support and my server admins, we determined MT was simply running out of memory on the server. My server admins increased the memory allocation to 100MB per process and all is fine.

It appears MT is working on 4.2 specifically to resolve some of these issues. Look for documentation on improvements in MT 4.2.

I have had that problem too after our server was being restructured and had stricter memory limits due to "server security" for the hoster.

Today I again talked to them, they put up the memory limit and nearly everything works fine again. There is still a 500 error, if the rebuilding process takes too long due to too many entries (individual entry archives).

Thanks for this hint, works fine.

Had the same difficulties and my host upped the limit, too. I'm not getting the 500 server error any more, but posting is PAINFULLY slow and takes forever. I occasionally get "double" posts, and triple comments from people, and then deleting the extra crap takes a million years, too. I'm basically just not posting anything because it's too big a pain.

Do you have LaunchBackgroundTasks enabled in your mt-config?

No clue whether I do or not. Does it help or hurt? Turn it on or off? Where do I turn it on or off?

Figured out that I definitely don't have launch background tasks in my config file, so that's not the issue.

OK. On some installs, with LaunchBackgroundTasks set to 1 (enabled) it could cause a problem.

Have you disabled the stats dashboard?

What plugins are you running?

Like a lot of other people, I've been having this problem a lot lately and got my account frozen when MT finally filled up my server space with 2 gigs of core dump files. I was able to sort that out with no problem, but I still get frequent 500 errors and core dumps.

Tonight, I got extremely frustrated as I tried to write a relatively long blog post, only to find that MT wouldn't save it. Note, I'm not publishing the post, just saving it.

So thinking that this was an issue with my MT 4.1 install, I tried upgrading to 4.2. Before doing so, I did the standard mysql and server backups (just in case), and I was able to use MT to publish a quick note on my blog that I was upgrading (no problems composing, saving, or publishing that post). I upgraded with no problems, updated my "upgrade" post with a note that upgrade seemed to be successful, and then went back to my entry.

Still no luck. I keep getting 500 error ("Premature end of script headers" etc...) and core dumps. I try adding portions of the post, little by little to narrow down the error, but there doesn't seem to be much rhyme or reason to it. I finally decided to look at one of the core dumps, and it's mostly gibberish, but towards the end of the file I see a ton of HTML... and not from my post or the MT interface - it's from Amazon.com. I have a bunch of amazon links in my post, so I try removing them and voila, everything works fine.

This, of course, makes no sense. Why does MT care about whether or not I have Amazon links in my post? Or is this some sort of spam thing my host is scanning for? Someone mentioned something like that in an earlier post in this thread, but it seems like a bit of a stretch to me.

Now I'm goint to try and finish up my post and publish it. Ever since I removed the Amazon links, I seem to be doing fine. Thought I should add this comment in case anyone else was having similar problems.

Generally speaking, MT doesn't care what's in your post content. As far as just the application, there's no reason Amazon links should do anything special.

You probably should talk to your host, though. The thing you said someone mentioned higher up was probably mod_security(I'm not about to trawl 11 pages.) The problem is that every host will configure it in their own way so there's absolutely no telling from the outside what might be causing the problem or, more importantly, why. You've got it narrowed down to Amazon links, but those tend to be long and ugly; there's all sorts of stuff in there. It might have nothing to do with Amazon at all. The real trigger could be some little string of letters. Here's an especially fun/weird/annoying example from a while back.

Send them a quick note describing what's going on, and the problem links themselves so that if this is what's doing on they can figure out why your links are triggering it and hopefully work around them.

The core dumps are likely something else altogether(a common reason is ImageMagick problems). That just plain shouldn't be happening. But try to clear up one problem first and see if it continues.

*** POSSIBLE FIX ***

I have had similar behavior since upgrading to 4.2. It got so bad that I had recreated the site and attempted to restore a backup. The problem followed the backed up data. I noticed that the restore hung during the assets import so I ended up creating a new blog and then export/imported the data into it.

A few days later I had the SAME problem with the new blog. After pulling my hair out for a while I decided to look at the asset tables closer. I didn't see anything out of the ordinary but I decided to clear the table for kicks. IT WORKED!!!

I talked to one of my authors and he said that he was having trouble uploading an image. I got a copy of it and sure enough, something with that image corrupts the asset tables and in turn the admin interface. As soon as I clear the asset entry for that image everything starts working again!

I have reported this to Sixapart as well...

Hope it helps...

P.S. - I helped another forum member clear his asset table and it fixed his problem too.

By "clearing your asset table" do you mean that you deleted all assets? Even if effective, this isn't really an ideal fix. I'd like to be able to post images on my blog.

I too am having problems with the assets table. After I upload a second image to a blog MT seizes up. It won't publish indexes. If I clear the asset from the MT-Assets table everything returns to normal, the image is there and shows in the posts, until I post another image and then problem repeats. Also, a click on the Manage Assets link seizes MT up and eventually results in 500 error.

I have no clue as to what is happening and can't seem to find a specific error. I do see unexpected end of headers, timed out before finishing, and out of memory error messages in the Apache error log.

Could this be related to Image::Magick Perlmagick? My mt-check says Image::Magick not installed (it is) and I can't seem to get that to install so MT sees it.


Update:
Clicking on Create:Asset also freezes MT - using Activity Monitor I can see that the _www user accessing perl is sucking up all the memory and jamming up the server.

I'm on MT 4.21, what I've figured is that I face this problem when uploading GIF files. Either from the MT dashboard or directly from custom field edited entry input. In order to resolve this issue, before publishing, you will need to delete all failed assets and entries from your system. Sounds really annoying, but I did not have problems with JPG and PNG files.

I'm not sure if this is an issue with MT (which I don't really don't wanna put time on upgrading) or my webhost's plugins

Any further assistant from Sixapart will be appreciated

I'm having this issue as well: 500 errors seemingly at random when I attempt to save an entry using MT 4.2.21. Sometimes it works, sometimes it doesn't. I've been having this problem since MT 4.0, and have done several clean installs in an attempt to fix it. The new installations work for a little while, but the problem inevitably crops up again. I am not going to waste my time trying to reinstall again until I have a guarantee it will work.

Instead I have done the following:

1. Set HTMLPerms to 0644 in mt-config.cgi per the instructions on this thread:
http://forums.sixapart.com/index.php?showt...p;mode=threaded

2. Cleared the contents of the mt_asset table in my MySQL database, per the suggestion of ratradon above.

Neither one has solved the problem.

Some interesting information from my MySQL runtime info, which may or may not be related (I don't know enough about MySQL to tell):

low_queries 92
Handler_read_rnd 300 M
Handler_read_rnd_next 3,649.46 M
Qcache_lowmem_prunes 14 M
Created_tmp_disk_tables 955 k
Select_full_join 228 k
Select_range_check 2,279
Sort_merge_passes 11 k
Opened_tables 2,402 k
Table_locks_waited 54 k

There's a lot more info in there--I just pulled these out because they are highlighted in red to indicate there is some kidn of problem. Like I said, I don't know enough about this to have a clear understanding of what any of these numbers mean. But there are a lot of suggestions to "optimize your queries" next to these numbers, as well as hints that tables may be "poorly indexed".

Any help with this or suggestions would be much appreciated!