Dear Squid Users,
We are using squid 2.2 stable 5 and I do not know how to analyse this kind of trouble we meet a couple of time per month :
-) The proxy may become very slow, because it starts using 98% of the CPU time (usualy, I am very lucky, it stays at about 7-10% all the time)
-) Users can't get any more pages. They always get a Broken Pipe error message for every sites
-) The folowing is what I have in my cache.log
-) Sometime, /etc/rc.d/init.d/squid restart (Redhat) is enought, but last time I had to reboot the server
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible. (US=UNAVAILABLE)
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (32) Relais brisé (pipe).
2000/06/13 10:25:59| sslReadClient: FD 17: read failure: (104) Connexion ré-initialisée par le correspondant (US=RESET BY PEER)
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:25:59| sslWriteServer: FD 21: write failure: (11) Ressource temporairement non disponible.
... a lot of these lignes and ...
2000/06/13 10:26:37| sslWriteServer: FD 27: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:26:37| sslWriteServer: FD 27: write failure: (11) Ressource temporairement non disponible.
2000/06/13 10:26:37| sslWriteServer: FD 27: write failure: (32) Relais brisé (pipe). (US=BROKEN PIPE)
Have you ever meet that ?
What can I do ?
Where have I to look into ?
Thanks.
Received on Wed Jun 14 2000 - 13:01:38 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:01 MST