Site Crashing
-
-
- John G. Walter
- Sr. Rocketeer
- Posts: 154
- Thanks: 0
-
I have this recurring problem with one of my sites crashing, displaying the following error message:
Table './jgw_live/zfdet_session' is marked as crashed and last (automatic?) repair failed SQL=INSERT INTO `zfdet_session` (`session_id`, `client_id`, `time`) VALUES ('ssdvdjim9iuvve6vlocud97kb0', 0, '1377004620')
Emptying the data out of that table in PHPMyAdmin is an easy fix, but I haven't figured out what the cause is?
At first, it happened when I installed a newer template (Stratos) onto my site, which was running Zephyr. However, I have restored from an older, supposedly stable version before that install, and now it is still happening.
My site is located on an Amazon EC2 Micro Instance, which I manage myself. I have several other sites there, all running without any problems.
What I have noticed, and I do not know if this is normal or not, is that when the site is up and running, looking at that table shows an added entry every few minutes. Yesterday, that quickly added up to over 900 entries, although later, many of them seemed to disappear, with the number down in the 600's.
This leads me to believe that something is looping in the site without my knowledge????
Any ideas?
Thanks.
-
-
-
- Reggie Simmons
- Preeminent Rocketeer
- Posts: 5200
- Thanks: 530
-
Hello John,
I've been told the issue involves client hitting or an issue on your Instance.
Nevertheless, our products have nothing to do with session management. - Kind regards,
Reggie
Learn to code for free.
www.freecodecamp.org
-
-
-
- John G. Walter
- Sr. Rocketeer
- Posts: 154
- Thanks: 0
-
Reggie Simmons wrote:
Hello John,
I've been told the issue involves client hitting or an issue on your Instance.
Nevertheless, our products have nothing to do with session management.
Thanks for the response. I realize it has nothing to do with RT products, but was hoping somebody might have some insight?
There isn't much traffic on this site, and the problem appears to be with excessive I/O requests with the EBS volume. Not sure why, as it just started on its own about a month ago or so, and a complete instance restore from a much older snapshot did not help.
-
-
-
- Reggie Simmons
- Preeminent Rocketeer
- Posts: 5200
- Thanks: 530
-
Hello John,
Unfortunately, I do not have any insight on this issue at this time.
Have you tried contacting Amazon for support? - Kind regards,
Reggie
Learn to code for free.
www.freecodecamp.org
-
-
-
- John G. Walter
- Sr. Rocketeer
- Posts: 154
- Thanks: 0
- Thank you anyway, for your responsiveness. FYI, I solved the issue. It appears that I had the session limit in Joomla set to 180 minutes, therefore 180 minutes worth of session logs were piling up before they were deleted. This hit the wall in my server's memory, which crashed the table. Shortening the session time-out to 30 minutes keeps the session table entries in the 250-400 range, instead of 3,000, and does not stress the memory. It didn't help, of course that I had 10 other sites on the same server instance.
-
-
-
- Reggie Simmons
- Preeminent Rocketeer
- Posts: 5200
- Thanks: 530
-
Hello John,
You're welcome! Thanks for sharing. - Kind regards,
Reggie
Learn to code for free.
www.freecodecamp.org
-
Time to create page: 0.066 seconds