[Check_mk (english)] Any way to adjust the 110 second timeout for distributed actions?
Antti.Makela at vintor.fi
Fri Feb 15 12:47:47 CET 2019
The timeout is hardcoded apparently in file
self._request_timeout = 110 # seconds
I'm not sure if changing this would have all the desired effects...
- Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
- Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -
From: checkmk-en <checkmk-en-bounces at lists.mathias-kettner.de> On Behalf Of Dr. Maik Götte
Sent: perjantai 15. helmikuuta 2019 12.28
To: checkmk-en at lists.mathias-kettner.de
Subject: Re: [Check_mk (english)] Any way to adjust the 110 second timeout for distributed actions?
We also experience these issues, additionally when renaming e.g.
switches. ASFAIK there is no such option anywhere; it seems hardcoded
somewhere. I played with timeout options in apache and in some of the
cmk python scripts without success. At the moment we have an open ticket
with cmk support on this.
Am 14.02.2019 um 13:11 schrieb Mäkelä, Antti:
> in some cases, when we do discovery for a new host on a distributed, slave node, the host responds slowly (especially across slow links). Thus, covering all the inventorizable MIBs takes a long time.
> However, the WATO times out after 110 seconds and cancels the process.
> Running cmk -I <host> from the slave node directly works fine.
> Same occurs also if we e.g. rename host that has a lot of graphing data etc to go through. We can see the cmk process still doing the find-replace operation when checked dirertly from slave node, but master does not like to wait.
> Where can we configure this 110 second "hard limit" for how long to wait for answer from distributed node? Even better would be if master would be aware that action is still ongoing even if it is taking a long time and give an *option* to cancel it if needed.
> Running 1.5.0p9.
More information about the checkmk-en