PS crash after starting enormous amount of processes

[mrtommy]

Hi

I would highly appreciate som tips and ideas how to continue fault tracing. We have PS 1.6.1 and run multistore mode (2 shops). We use WT_choicemarket theme and another theme specialdev_bootrap somehing on the secondayr site.

 Also we only run original module for PS and choicemarket theme module (WT). The server has 512MB memory. We have installed the shops and swithed to PS 1 month ago and until a couple of days ago everything run smooth I would say... we have until now used smarty cache.

Our shops contains aorund 14K products and 40K customers...

 

What happens:

- with good help from the technical team at the hosting company Binero we can say that PS start up away too many processess and crash the site. They dont do application fault tracing and cannot see wich PS processes

- It typically happens on one shop at the time or sometimes both shops. It could keep the site down between 5min and 1 hour. And the sites are seldom working longer than 3-5 minutes before the processes hits the roof and the server is not reachable.

- Binero says that this is not a memory issue or lack of resources at server side, we have strong power and low amount of visitors. They are sure something in PS application is doing loops or not shut down executed processes (or similar).

- When the problem occurs on a shop we sometimes get white page server error message (resources reach limit, come back later..). But mostly the browser is just chewing forever and nothing happens when refreshing. One time we got a PS error showed up saying (one time only) uncaught exception . ANd 3-4 times we have seen a server error white page with text "premature end of script headers: index.php"

- We cannot reach BO or Frtontypage until the processes calmed down under limit level. THe way for us to speed up the time to get into BO and do adjustment is to deny IP accept ours in .htaccess

 

 

Could it be something with that modules are uninstalled but hooked in "positions"....

We have run out of ideas, please help brainstorm....

 

Yesterday we did a trick that helps the sites stay barely just on the right side of limit. We shut down the smarty cache totally. Now the technicans see an improvment and the sites are continously reachable (but a bit slower). The thing is that we still are away to close the limit and the support people att binero are clearly stating that we still have the issue....

 

Please help

Thanks in advance

Edited October 28, 2015 by mrtommy (see edit history)