https://bz.apache.org/bugzilla/show_bug.cgi?id=64306

--- Comment #4 from [email protected] <[email protected]> ---
On May 1:

May 01 01:04:24 joooj ocsp_proxy[715]: [info] connection from ::1:38408
May 01 01:04:29 joooj ocsp_proxy[715]: [debug] got request
May 01 01:04:34 joooj ocsp_proxy[715]: [debug] cache key is
ocspxy_b390a7d8c9af4ecd613c9f7cad5d7f41fd6930ea0x4aa43b32a0f5e7b491c2881d02818ca5
May 01 01:04:35 joooj ocsp_proxy[715]: [debug] sending response
May 01 01:04:36 joooj ocsp_proxy[715]: [debug] disconnecting ::1:38408

I've eventually found with atop that ocsp_proxy was almost entirely swapped out
shortly before (from 11 MB RSIZE to 3.6 MB RSIZE), and the cause was that I, as
a normal user, had run a process that needed much memory. And it was back to 11
MB after the "disconnecting".

Given the fact that the kernel can swap out a small system process like
ocsp_proxy because a non-system process needs much memory, I think that apache
should take such a case into account, i.e. increase the default timeout. A
8-second timeout seems very low anyway, in particular for something that would
yield an error and make the server unavailable (due to bug 57121). One
generally sees timeouts from 1 to 3 minutes. And I don't see why a timeout
value larger than 8 seconds should not be used here.

-- 
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to