We're seeing instances where Prince appears to be bringing down our server (RedHat 5.6 2). Prince 7.1 is used as the rendering component in a web-based typesetting system. To quote the IT department:
At that time the server had 2 2.67GHz proccessors and 4GB of RAM (5333.52 bogomips per processor). We then added more CPUs and Memory, to 4 CPUs and 8GB RAM. Less than one hour later, the prince processes had used up all the available RAM and swap space.
I'm not a server guy, so I don't really understand all this. I have looked through the Prince log file, and haven't seen much that looks unusual, aside from:
(but that message has only appeared twice, and doesn't seem to be correlated with the crashes).
Any advice on helping to track down the problems would be greatly appreciated. Note that the volumes seem fairly high; Prince could be called thousands of times a day.
Thanks!
Dave
What we’ve been seeing is that the 'prince' process starts eating up RAM and swapping like mad. We have to kill it off and restart httpd and instantly it uses several gigs less RAM.
At that time the server had 2 2.67GHz proccessors and 4GB of RAM (5333.52 bogomips per processor). We then added more CPUs and Memory, to 4 CPUs and 8GB RAM. Less than one hour later, the prince processes had used up all the available RAM and swap space.
I'm not a server guy, so I don't really understand all this. I have looked through the Prince log file, and haven't seen much that looks unusual, aside from:
Wed May 11 17:46:57 2011: internal error: pack_blocks: empty list(but that message has only appeared twice, and doesn't seem to be correlated with the crashes).
Any advice on helping to track down the problems would be greatly appreciated. Note that the volumes seem fairly high; Prince could be called thousands of times a day.
Thanks!
Dave