DEQUEUE_ERR - Error querying XMLRPC Server.

Hi!

Videocache rotate its logs automatically but it doesn't rotate squid's logs. Squid log rotation is known to affect videocache functionality especially the XMLRPC part. There are three solutions.

1. You stop automatic rotation of squid logs and rotate them manually at low traffic times. But this may be painful if you have to do it a lot of times.

2. [Not Tested] Add a line '/etc/init.d/squid reload' or 'service squid reload' to /etc/logrotate.d/squid file after squid rotates its logs i.e. after 'squid -k rotate' line. This may work.

3. Restart squid after log rotation. It is similar to point 2 with the difference that instead of adding the line to reload squid, you'll have to add 'service squid restart' command to /etc/logrotate.d/squid file.

Thank You!

hi,

i did a manual log rotation "squid -k rotate" and i experienced the error XMLRPC. now i know it will affect videocache functionality...during the trouble i did kill all videocache.py's and restart squid. then no more error.

is it ok if i remove the file /etc/logrotate.d/squid? i will just run squid -k rotate manually then. after rotating squid logs, will restarting squid removes the XMLRPC error?

thanks,

vincent

Hi Vincent,

That should be fine I think. If you are comfortable in rotating the squid logs manually once in a while, /etc/logrotate.d/squid is not needed.

The solution you explain will work without any problem.

If you still face issues, please let us know in forums :)

Thank You!

Hi Vincent!

That normally occurs when either XMLRPC server has gone away or it stops responding. Squid reload should do the job. Update-vc doesn't help in restarting squid or videocache.

Thank You!

Hi Vincent,

What is the avg load on your servers and how many clients do you have?

Thank You!

hi kulbir,

avg load of squid = 5%
memory used by squid = 5%

now, i have a new problem, my set up is transparent proxy, when i turn on transparent proxy, it will become very slow...until u could hardly open a browser but no RTO when u ping to any wbsite.
my videocache directory reached to 80gigs already. i have an active clients right now of about 150 clients...yesterday it was very okay even the connected clients were 350 subs...im confused what could be causing this. i turned off the squid right now. im still looking for reasons why it gets so slow when i turn own the squid in transparent mode....please advise...

thanks,

vincent

is it the videocache slowing my connection? for 1 week time, the directory of the videocache reached 80gigs already for almost 400 clients connected with average active numbers of 200, maybe.

thanks,

vincent

Hi Vincent,

Videocache need not be configured in different ways for transparent or non-transparent proxies. Please mail me your config files for squid and videocache.

Thank You!

Hi Kulbir!



I made changes right away after receiving your reply mail. I monitored my server, the number of active clients reached to 200 at around 9pm. By 12MN, they’re only 185 and slowly decreases. But, by 1AM, my videocache.log again got empty  so I restarted squid. Im afraid il have to ask again why this is happening. Also, is there any means of limiting http request of clients. Because maybe lastime my squid got flooded by http request of a single client. Could it be done using iptables considering my setup as transparent proxy?

Thanks,

Vincent