> On 30 Aug 2019, at 16.40, Alex Rousskov
> wrote:
>
> On 8/30/19 8:16 AM, Ilari Laitinen wrote:
>
>> I suspect Squid might be waiting for local TCP ports from the kernel
>> (or something related).
>
> IIRC, ephemeral source port allocator is instantaneous -- Squid either
> gets a port or a po
On 8/30/19 8:16 AM, Ilari Laitinen wrote:
> I also noticed that the platform (in general) tries to resolve ipv6
> first, but the TCP dumps have no ipv6 packages at all. This is
> baffling, because there were indeed some unrelated open ipv6
> connections on the Squid server (reported by netstat).
>> I noticed small but consistent spikes in squid's disk cache usage
>> coinciding with the issue at hand. This seemed strange, given there
>> was no other traffic during the tests and proxied HTTPS means there's
>> nothing to cache (right?).
>
> Correct. To avoid suspecting disks, configure Squid
On 8/20/19 9:12 AM, Ilari Laitinen wrote:
> I am experiencing a slowdown between CONNECT requests and corresponding "200
> Connection established" responses.
I assume you are not using any SslBump features but please correct me if
I am wrong.
> source —> squid [SYN]
> source <— squid [SYN, ACK
Hi!
I am experiencing a slowdown between CONNECT requests and corresponding "200
Connection established" responses. This happens when performance testing a
load-balanced pair of squid servers with hundreds of simultaneous, independent,
proxied HTTPS requests per second per server.
Most new con
