On 1/18/2013 10:48 PM, gardner@ns1.forge.name wrote:
Are you seeing core dumps?
Yes (at least, on this last crash, there were several dumps).
Are you tracking the number of threads running?
I'm not certain how to do that. One of my issues is that when it freezes, I can't get any information. The only way I've found to get it back online is to perform a hard reset. All performance stats when the server is running fine appear to be very low impact.
Are you looking to see how much memory is used by ppid?
I had hoped that Sysstat would give me this information over time, but when I needed it, that particular report was corrupt, and unreadable (ie. useless).
Look at your logs. What do you see on start-up?
Looks normal.
What about when you stop apache... anything interesting in the logs there?
Nothing that looks interesting to me.
Is it just something as simple as running out of drive space?
I've checked, and no, that's not it.
Have you trimmed down your various http.conf files to only what you need and ditch a bunch of modules you don't need?
I'm unclear on exactly what I need and what I don't, so no, I haven't.
Is it a version of apache that just has a bug in it? Do you need to build your own or possibly if you have built your own do you need to use a "stock" apache?
How would I find that out? I'm using the CentOS stock Apache now.
There are just so many things that can be done.
I'm very willing to look into this with you. Hit me up via my work email: jeffrey.gardner at sprint dot com
Please let me know when you are available. My boss is anxious to get these resolved, and thus pacify our client.
I really appreciate your willingness to assist me in this.
~ jay