Search squid archive

Re: Re: can't failover in round-robin

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Ken Peng wrote:
Hello again,

I watched cache.log and found these info:

2009/06/26 14:04:36| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:05:07| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:05:18| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:05:48| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:05:48| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:05:49| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:06:52| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:07:42| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:07:43| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:07:43| TCP connection to 192.168.1.101/80 failed
2009/06/26 14:07:43| Detected DEAD Parent: portsw_1_2


After long time (it's about  3 minutes), Squid found that peer was dead.
So my question is, by what way squid will find a peer is dead or alive?

By 10 consecutive request failures.

Why it takes so long time?

Because its 10* request timeout.

There are fixes in more recent Squid to improve this situation by making the number of retries configurable.



Thanks again.

Regards.



Dear Gurus,

I have configured squid-3.0.stable15 with these rules for an
accelerator:

cache_peer 192.168.1.100 parent 80 0 no-query front-end-https=auto
originserver name=portsw_1_1 round-robin
cache_peer 192.168.1.101 parent 80 0 no-query front-end-https=auto
originserver name=portsw_1_2 round-robin

acl port_1 myport 80
cache_peer_access portsw_1_1 allow port_1
cache_peer_access portsw_1_2 allow port_1


All run fine when both backend servers are fine.

But, when I shutdown one of the backend servers (for example,shutdown
192.168.1.101), I then accessed to Squid, got an error page and can't
open the website correctly.

The cache.log said:

2009/06/26 12:50:02| TCP connection to 192.168.1.101/80 failed


Why Squid doesn't failover between the two backend servers?
When 1.101 is down, squid should choose 1.100 as original server, and
should not show an error page to users.

Please help, thank you a lot in advance.

Regards.



Amos
--
Please be using
  Current Stable Squid 2.7.STABLE6 or 3.0.STABLE16
  Current Beta Squid 3.1.0.8

[Index of Archives]     [Linux Audio Users]     [Samba]     [Big List of Linux Books]     [Linux USB]     [Yosemite News]

  Powered by Linux