[nagios@nagiosprodxi1 libexec]$ /usr/bin/check_gearman -V
check_gearman: version 1.4_nagios4 running on libgearman 0.25
I would like to know how long does the Gearman collectors retain the monitoring data if there is connectivity problem with the Nagios XI host?
Does gearman collectors send all the collected monitoring data that is pending submission once connectivity is Ok?
I believe it holds a socket connection open until the timeout is either cleared, or reaches it's max elapsed time. Once either of these are met, the data is released. I can talk to some co-workers that may know a bit more about it's protocols when it comes to this, but as far as I know, there is not a lot behind the scenes, it's either sent to Nagios, or dropped when the timeout int is reached, I don't think there is any caching.
I do not believe gearman "caches" check results. As it is a distributed setup, if one worker goes down and the checks distributed to it are not returned, the master gearman daemon will just reschedule the checks on another worker.
Former Nagios employee
"It is turtles. All. The. Way. Down. . . .and maybe an elephant or two."
VI VI VI - The editor of the Beast!
Come to the Dark Side.
abrist wrote:I do not believe gearman "caches" check results. As it is a distributed setup, if one worker goes down and the checks distributed to it are not returned, the master gearman daemon will just reschedule the checks on another worker.
The collector will be at a remote site across the WAN.
So if there is any connection issue with between the collector and the Nagios XI host all checks will timeout and effectively all monitoring will be down for the site, right?
Essentially, yes, just as if you were checking a remote host by making a direct connection to an NRPE agent, and if you lost connectivity there, those checks would start timing out until the connection resumed.
If you can't connect to an address properly, monitoring will not be possible until a connection is made.
slansing wrote:Essentially, yes, just as if you were checking a remote host by making a direct connection to an NRPE agent, and if you lost connectivity there, those checks would start timing out until the connection resumed.
If you can't connect to an address properly, monitoring will not be possible until a connection is made.
Is there a way to setup an independent XI in a remote site which syncs its data to the Central XI?
You can follow a couple short documents to set it up on both servers. The XI system in that network can run all of that network's checks itself, and the simply mirror and push the results up to the central XI server.
You can follow a couple short documents to set it up on both servers. The XI system in that network can run all of that network's checks itself, and the simply mirror and push the results up to the central XI server.
Some Questions
1) Can I confirm multiple internal servers can sync to the central server?
2) Where does the alerts notifications get generated? The internal or central server?
3) Does this work with Nagios Core internal servers as well?
4) Any potential issues with this type of setup?
rajasegar wrote:1) Can I confirm multiple internal servers can sync to the central server?
We need to clarify this. Pushing checks to a central server will not "sync". They are literally pushed. If the central server is not reachable for a period of time, any checks run by a remote XI server will not get pushed to the central server.
Former Nagios employee
"It is turtles. All. The. Way. Down. . . .and maybe an elephant or two."
VI VI VI - The editor of the Beast!
Come to the Dark Side.
rajasegar wrote:1) Can I confirm multiple internal servers can sync to the central server?
We need to clarify this. Pushing checks to a central server will not "sync". They are literally pushed. If the central server is not reachable for a period of time, any checks run by a remote XI server will not get pushed to the central server.
I am looking to sync (update) the check data to the central XI.
Remote server can decide not to push anymore? It should be able to resume when connection is established.