We will be installing NagioXI on new hardware so we have the advantage of starting with a clean system. We are wondering if we should be looking to convert from NSClient to NCPA since NCPA is actively supported by Nagios. Looking for recommendations on how we might go about this? We would need to install the agent on about 950 window servers and 80 Linux servers. Also we would need to keep the current production environment up and running until the cut-over date. This would involve being able to test the NCPA set up while the NSCLient on the old production continues to run. Will this create a problem? We wouldn't keep the new environment doing checks 24x7, just during the day to verify they are working.
Any thoughts would be appreciated..
Greg
Converting from NSClient to NCPA
Re: Converting from NSClient to NCPA
Hi,
That seems like a great idea.
You would need to push out an NCPA install with your customized ncpa.cnf file and any
plugins you wanted to use to all your servers.
On all of your servers you would need to make sure tcp port 5693 is open for the new Nagios xi server.
Next you should configure 10 Windows and 10 Linux machines on the new Nagios xi server. Verify that
the new XI server is monitoring what the old XI server was monitoring. Let them run in parallel
until you are comfortable with how the new XI server is working. Be aware that if one of your
servers causes alerts that two notifications will be sent out - one from the old XI server and one
from the new XI server.
Once you have things set up the way you like then you can proceed to remove the
10 Windows and 10 Linux servers from the old XI server.
Now you can migrate the other 940 and 70 servers from the old to the new XI server. This document will
help: https://assets.nagios.com/downloads/nag ... gement.pdf
My preference for doing this is to do it in manageable chunks so it is easy to recover from any hiccups.
Once everything is migrated you can clean-up any unused software (NSClient).
Thanks
That seems like a great idea.
You would need to push out an NCPA install with your customized ncpa.cnf file and any
plugins you wanted to use to all your servers.
On all of your servers you would need to make sure tcp port 5693 is open for the new Nagios xi server.
Next you should configure 10 Windows and 10 Linux machines on the new Nagios xi server. Verify that
the new XI server is monitoring what the old XI server was monitoring. Let them run in parallel
until you are comfortable with how the new XI server is working. Be aware that if one of your
servers causes alerts that two notifications will be sent out - one from the old XI server and one
from the new XI server.
Once you have things set up the way you like then you can proceed to remove the
10 Windows and 10 Linux servers from the old XI server.
Now you can migrate the other 940 and 70 servers from the old to the new XI server. This document will
help: https://assets.nagios.com/downloads/nag ... gement.pdf
My preference for doing this is to do it in manageable chunks so it is easy to recover from any hiccups.
Once everything is migrated you can clean-up any unused software (NSClient).
Thanks
Re: Converting from NSClient to NCPA
Thanks.
Is it possible with NCPA to do a single call for disk usage if you have multiple disks, and do a showfail. Whether there is one disk or 5 with one call you check them all and only show the one that fails?
Is it possible with NCPA to do a single call for disk usage if you have multiple disks, and do a showfail. Whether there is one disk or 5 with one call you check them all and only show the one that fails?
Re: Converting from NSClient to NCPA
Hi,
When you say "show the one that fails" do you mean in the Service Status dashboards? Or do you mean
you want to get an alert that specifies which disk is disk is failing if you have say 5 disks?
Thanks
When you say "show the one that fails" do you mean in the Service Status dashboards? Or do you mean
you want to get an alert that specifies which disk is disk is failing if you have say 5 disks?
Thanks
Re: Converting from NSClient to NCPA
With NSClient we do a 'Showfail' and the one that is critical will be what is displayed in the alert/notification.
Re: Converting from NSClient to NCPA
I'm sorry, I stated the wrong thing in the previous post. It is not showfail.
In one command we are looking at all drives on a window server and seeing which ones have reached the warning or critical threshold. This is a more efficient way and will catch any new disks that get added
When we had looked at NCPA a while ago you would need to list each disk separately and if a new disk was added it could be missed.
Greg
In one command we are looking at all drives on a window server and seeing which ones have reached the warning or critical threshold. This is a more efficient way and will catch any new disks that get added
When we had looked at NCPA a while ago you would need to list each disk separately and if a new disk was added it could be missed.
Greg
Re: Converting from NSClient to NCPA
Hi Greg,
"When we had looked at NCPA a while ago you would need to list each disk separately and if a new disk was added it could be missed."
Yes - I verified that on my systems here.
Each disk having its own monitor is pretty typical. The problem with checking multiple disks with one check
is that if you are checking a disk that is currently present (and returning performance data for it), if the disk ever gets
removed the performance data for all the drives in that check will stop updating because the RRD is expecting
that additional datasource (the removed disk) when inserting data.
When you try to insert more or less datasources into an RRD it will fail.
That is the reasoning, but it probably doesn't help you. When using NCPA it will perform discovery upon the first configuration
connection it makes, but there is no provision for it to attempt discovery again. I looked in Nagios Exchange, and nothing caught
my eye, but you could take a look.
One idea would be to run a script via cron to check for new/deleted disks. Then you could use the api to edit the config to
account for the change.
Thanks
"When we had looked at NCPA a while ago you would need to list each disk separately and if a new disk was added it could be missed."
Yes - I verified that on my systems here.
Each disk having its own monitor is pretty typical. The problem with checking multiple disks with one check
is that if you are checking a disk that is currently present (and returning performance data for it), if the disk ever gets
removed the performance data for all the drives in that check will stop updating because the RRD is expecting
that additional datasource (the removed disk) when inserting data.
When you try to insert more or less datasources into an RRD it will fail.
That is the reasoning, but it probably doesn't help you. When using NCPA it will perform discovery upon the first configuration
connection it makes, but there is no provision for it to attempt discovery again. I looked in Nagios Exchange, and nothing caught
my eye, but you could take a look.
One idea would be to run a script via cron to check for new/deleted disks. Then you could use the api to edit the config to
account for the change.
Thanks