Support Queries

Scheduler not working

Hello!

Today I had to reboot my proxy server because at videocache.log I saw that any video was scheduled for caching.

Wich can be the cause?

Using version 1.7.

Thanks,

Josep Pujadas

8 Answers

Hi Josep,

You need not reboot your system in case of inconsistencies. You can just restart the squid and python processes and doing that will restore the sanity of squid and videocache and everything will be fine. Its a program in development, scheduler may go off because of XMLRPC server. Thats the only cause I can think of right now.

Thank you for reporting. But I would request you to upgrade to 1.8 as soon as possible.

Kulbir,

Ok, I will migrate from 1.7 to 1.8 a.s.a.p.

However, how can I test if XMLRPC server is working?

Regards,

Josep Pujadas

Josep,

All you need is just one command

[root@localhost ~]# netstat -atpn |grep 9100

If its bound and someone is listening on this port, its almost sure that XMLRPC server is working.

Thank You!

Kulbir,

Ok, thanks!

In my case (FreeBSD), this worked:

netstat -an | grep 9100

Regards,

Josep Pujadas

Hi Kulbir,

On my test server, after sitting there for the night without being used, I am seeing a high CPU usage, running top I see an average of 7.00 for the las few hours and I see 7 python PID's

ID      Owner       CPU     Command   
15534   squid   21.0 %  (python) /usr/share/videocache/videocache.py
15530   squid   20.6 %  (python) /usr/share/videocache/videocache.py
15536   squid   20.3 %  (python) /usr/share/videocache/videocache.py
15533   squid   20.1 %  (python) /usr/share/videocache/videocache.py
15532   squid   20.0 %  (python) /usr/share/videocache/videocache.py
15531   squid   19.8 %  (python) /usr/share/videocache/videocache.py
15535   squid   19.8 %  (python) /usr/share/videocache/videocache.py
15527   squid   1.9 %   (squid) -D -f /etc/squid/squid.conf
19176   root    0.8 %   /usr/libexec/webmin/proc/index_cpu.cgi
4129    root    0.2 %   top

What can be causing this?, and how can I avoid it? If this happens on a production it will put the system on it's knees... please point me in the right direction on how to fix it....

Regards from Mexico

J. Boardman

Jorge,

It may be caused due to incorrect permissions of the log files or cache directories.

BTW, what is approx load (users and requests per second) on your proxy machine?

Thank You!

Hi, thanks for answering,

I think my permissions are OK, this occurs sometimes; sometimes not.... weird. (Please tell me if something got to be different from default user "squid")

On the BTW question, Nobody is yet using this server except me, it is only for testing purposes but planned to go into production next week, it will serve around 200-300 users, this server is a Dell Poweredge SC1425, Dual Xeon 3.0Ghz, 6GB RAM, 320GB Sata Disks... and it is only running Squid and VideoCache on Fedora 10, it should be good enough I may think...

But that occasional issue on leaving several python processes hanging around will definitely put the server on it's knees real quick. (This occurs during the night nobody using it)

Any other ideas, please?

JB

Jorge,

The most expensive part of videocache (computationally) is the cache directory size calculator which is run frequently. It may really eat up all your CPU. And this component is really tightly coupled with videocache.

Contact me using contact tab, if you are interested in special modification.

Thank You!