I am running Ubuntu 16.04 with a Squid on cloud server.
Squid is used for my Android application to access few web sites.
In general everything worked fine until 3 days ago during the highest load period (9 PM - 11 PM), when I have ~20 000 active users. During this time I am getting a lot of Squid logs like this:
2017.09.13/19:08:46 0 79.143.34.171 TAG_NONE/503 0 CONNECT login.vk.com:443 - HIER_NONE/- -I'm investigating already for 2 days, and what I've tried so far is disable caching, disable IPv6, using via off and forwarded_for off, using http_port 0.0.0.0:8080 to force IP v4, setting tcp_outgoing_address <server_ip_address>, and some more things, that I currently can't remember.
So firstly I don't understand that 503 code: is that the response from the web site login.vk.com, right? Does it mean that that site (server) is overloaded by requests, or just exceeds the requests per second from my server's IP?
How can I fix this, or at least improve the situation? Maybe I should somehow limit max simultaneous connections to my Squid?
1 Answer
I had the same problem with a web page of a bank, the log is:
TAG_NONE / 503 0 CONNECT enlinea.bt.gob.ve:443 - HIER_NONE / - -What I did was place:
pinger_enable off
half_closed_clients off
quick_abort_min 0 KB
quick_abort_max 0 KB
quick_abort_pct 95And change:
client_persistent_connections on
server_persistent_connections onBy:
client_persistent_connections off
server_persistent_connections offIt immediately began to work after changing those settings.
2