1

I am having a problem with some of my Magento websites. I get the message "The connection was reset" in Firefox or "No data received" in Chrome, while my other sites (custom build PHP sites) are working perfectly. This happens time by time (for example once in a while, like 4 to 5 days)

I have searched for apache log error but there is nothing strange there.

I have been using APC for cache and i tried without it but after some time i get this error again and it is solved only by restarting the Apache server.

Environment information: Linux 2.6.32-431.el6.x86_64 #1 SMP Fri Nov 22 03:15:09 UTC 2013 x86_64 running CentOS PHP Hander: DSO (because i am using APC) PHP Version 5.4.24 Apache 2.4 (already had the 2.2 but the same problem) Intel® Core™ i7-920 Quad-Core 48 GB DDR3 RAM

UPDATE: This problem happens only with Magento 1.7, i have and a 1.8 Magento site in the same server but it works fine when the other site has this problem.

I don't really understand what can cause this problem.

Teja Bhagavan Kollepara
  • 3,816
  • 5
  • 32
  • 69
user1305626
  • 123
  • 1
  • 5
  • Can you check what are the processes related to PHP running when you get the error? – Emi Feb 25 '14 at 14:34
  • @Emi i have to wait for the error to occur again :/ – user1305626 Feb 25 '14 at 14:36
  • 1
    The error might be caused by the fact that some cron is running, that would set locks on the mysql tables. So you should look for the running processes and then see what they are doing. – Emi Feb 25 '14 at 14:46
  • what is your environment? which versions of php,apache, which OS, do you use php as apache module, via cgi or via fcgi? if fcgi, managed by a apache module or php own solution?

    Is there anything in the apache error logs? For example a segfault?

    – Flyingmana Feb 25 '14 at 16:23
  • If a restart of Apache fixes your issue, the culprit is absolutely Apache. Check to make sure your config settings are correct. It should also throw a log in your Apache logs at the exact time of the event. To me it sounds like a timeout/termination is happening to an open/running Apache process. – Mark Shust at M.academy Feb 25 '14 at 20:36
  • @Flyingmana i just updated the environment specification. – user1305626 Feb 26 '14 at 08:32
  • @MarkShust i thought so but the other Magento 1.8 sill keeps working after the 1.7 version stops and shows these messages. In the apache log file there is no any error information in the time that the problem is occurred. – user1305626 Feb 26 '14 at 08:34
  • 1
    If nothing is in the logs, there is either no problem or not enough logging. Iam pretty sure, you should have a segfault error in some of the logs.

    Are you sure without APC the Problem still exists? Anyway, the combination APC+PHP5.4 is dangerous, as it is unstable. You should update to PHP5.5 as soon as possible or look how to install the ZendOpCache for 5.4

    – Flyingmana Feb 26 '14 at 09:22
  • @Flyingmana it still happens after i remove the APC configuration from Magento, now i just disabled APC at all form php.ini so it's not running. The PHP 5.5.9 is marked as experimental in WHM that's why i have not installed it. I am using WHM and i am checking error logs of apache in /usr/local/apache/logs error_logs and /usr/local/apache/domlogs for the specified domain. I will have a look again. – user1305626 Feb 26 '14 at 09:38
  • WHM -- are you running on a shared host? Most likely you are running out of memory if that is the case. If so, I'd get on a decent hosting account first then go from there. – Mark Shust at M.academy Feb 26 '14 at 12:28
  • 1
    @MarkShust as mentioned in the post the server is dedicated and it has Intel® Core™ i7-920 Quad-Core 48 GB DDR3 RAM so i don't think that could be the problem – user1305626 Feb 26 '14 at 20:26
  • Have a look through http://magento.stackexchange.com/questions/428/fundamentals-for-debugging-a-magento-store/429#429 and see if any of these debugging tips help you find out what is going wrong. – David Manners Mar 01 '15 at 12:31

1 Answers1

0

Set in the php.ini file:

error_reporting = E_ALL

Then restart Apache. If you are getting a blank page, the error should definitely be logged.