Hi Henry, Please have a look at https://www.varnish-cache.org/docs/2.1/tutorial/advanced_backend_servers.html#health-checks. Here you can explicitly configure how Varnish will check the health of each backend.
Lars -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Henry M. Umansky Sent: Thursday, August 18, 2011 4:41 PM To: [email protected] Subject: HTTP 503 Service Unavailable errors Hello, I recently installed Varnish from the Redhat EPEL repository (Varnish 2.1.5-1.el6.x86_64) and I'm having a slight problem with load balancing. I currently have two backends: director default round-robin { { .backend = { .host = "webhead01.example.com"; .port = "http"; .connect_timeout = 1s; .first_byte_timeout = 5s; .between_bytes_timeout = 2s; } } { .backend = { .host = "webhead02.example.com"; .port = "http"; .connect_timeout = 1s; .first_byte_timeout = 5s; .between_bytes_timeout = 2s; } } } and when I shut down Apache on one of the backends and restart Varnish, I get "503 Service Unavailable" errors on random files (usually just one or two files). I never get this error when both backends are running. How does the Varnish probes work, do they ping the backend server to see if they are alive or actually attempt to connect to port 80 to see if Apache is alive? I'm relatively new to Varnish, so I want to apologize if I don't have all the information or missing a configuration somewhere. Henry Umansky Web Development Services Princeton University [email protected] 609-258-1674 _______________________________________________ varnish-misc mailing list [email protected] https://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc _______________________________________________ varnish-misc mailing list [email protected] https://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc
