Very likely a duplicate of
https://bz.apache.org/bugzilla/show_bug.cgi?id=63325
(regression in 2.4.39) with the fix
https://svn.apache.org/viewvc?view=revision&revision=1859376
which will be part of the next version 2.4.40.
You can try that patch if you like. 2.4.40 is not far away, maybe next week.
It is not yet in the published Changelog, because that is for 2.4.39 and
this change will be in 2.4.40. The new item will be:
*) mod_reqtimeout: Fix default rates missing (not applied) in 2.4.39.
PR 63325. [Yann Ylavic]
Regards,
Rainer
Am 01.08.2019 um 02:39 schrieb Charles Parker:
A recent change to mod_reqtimeout (2.4.39) is causing large uploads to
sporadically fail on my production server with a client error:
|ajp_read_into_msg_buff::jk_ajp_common.c (1553): (main_ajp13_worker)
receiving data from client failed. Connection aborted or network problems|
This is happening partway though chunked PUT requests to mod_jk.
From the changelog:
*) mod_reqtimeout: Allow to configure (TLS-)handshake timeouts.
PR 61310. [Yann Ylavic]
My short-term workarounds for this problem are:
a. Downgrade to httpd-2.4.38 (not recommended, see changelog for 7 CVEs)
b. Comment out the LoadModule line for mod_reqtimeout
For details, please see:
https://serverfault.com/q/972680/293818
Questions:
1. Was there any accompanying documentation for this change?
2. What configuration options are available to adjust this timeout?
3. Why does this result in a client error instead of a timeout error?
4. Has anyone else experienced this problem?
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
For additional commands, e-mail: users-h...@httpd.apache.org
