Werk #10245: Fix thread leak triggered by unstable site connections
Component | Livestatus proxy |
Title | Fix thread leak triggered by unstable site connections |
Date | Oct 1, 2019 |
Checkmk Edition | Checkmk Enterprise (CEE) |
Checkmk Version | 1.6.0p3 2.0.0i1 |
Level | Prominent Change |
Class | Bug Fix |
Compatibility | Compatible - no manual interaction needed |
The livestatus proxy normally keeps open several connections to the remote livestatus server and uses them in the moment a livestatus query is made from the central site to the remote site.
In the moment a site is not reachable anymore, for example because of a network connection issue, this is detected by one of these connections. The site is then terminating all other connections and trying to recover from this situation by continuously reconnecting to the remote site.
In this situation the livestatus proxy was not cleaning up it's old connection threads while opening new ones for the new connections. After some time this could result in "Resource temporarily unavailable" errors in various places or "can't start new thread" shown in the GUI.