When will NSClient ++ 0.4.x be provided ...

[PhilG]

in the Windows Server monitoring wizard stable client to download? It still provides 0.3.9.

I identified this while working on a different issue that I'm working on:
"NSClient - ERROR: Could not get data for 5 perhaps we don't collect data this far back?"
I have found a possible solution already for that error and will try to rebuild the counters.

[jdalrymple]

NSClient's code is really volatile right now and will require major testing before upgraded versions are linked in the wizards.

Have you noticed any breakages in the wizard using the newer client versions? Feedback is very useful but we have to do quite a bit of internal testing ourselves. The hope is that eventually we replace nsclient++ with our own ncpa that we can more easily keep current since the code will be developed in-house.

That doesn't answer your question I know, but unfortunately there is no official answer due to the dynamic nature of the nsclient++ agent. Internally we're all looking forward to leaving it behind in favor of ncpa.

[Box293]

NSClient 0.4.1.105 is stable.

[WillemDH]

I can confirm NSClient 0.4.1.105 is very stable. Have it running on 500 servers.

Grtz

Willem

[jdalrymple]

I can put in a feature request. I guess the things that we're most concerned about re: "stability" is that our builtin defaults in the wizard *just work* and additionally that our documentation:

http://assets.nagios.com/downloads/nagi ... ios-XI.pdf

Still coincides with all the installation steps and such without too much confusion for rookie users of the product.

Can you guys (WillemDH, Box293) verify those to be the case? This is the kind of testing I was referring to us conducting before updating the package in the XI wizard.

Also PhilG - worth mentioning: These are just hyperlinks to external packages. Additionally the link underneath the packages should take you straight to the nsclient++ download page which of course will have the most current (stable and beta) downloads.

[rajasegar]

NSClient 0.4.1.105 is stable.

Agreed except on Windows XP where it just stalls with Windows Semaphore error quite often.
Had to write a script to check the log and restart the service using Task Scheduler.

In fact all the version I tested so far until 0.4.1.105 has this issue with Windows XP.
The other OS versions no problem at all.

[lmiltchev]

I realize that some customers could be still using Windows XP but I believe at one point we will need to move on. It's support ended on April 8, 2014.

http://windows.microsoft.com/en-us/windows/lifecycle

[rajasegar]

I realize that some customers could be still using Windows XP but I believe at one point we will need to move on. It's support ended on April 8, 2014.

http://windows.microsoft.com/en-us/windows/lifecycle

Yeah, but in real life it is never the case.
Just a note in case people are having issues with the service.

[Box293]

in the Windows Server monitoring wizard stable client to download? It still provides 0.3.9.

I identified this while working on a different issue that I'm working on:
"NSClient - ERROR: Could not get data for 5 perhaps we don't collect data this far back?"
I have found a possible solution already for that error and will try to rebuild the counters.

Is this error related to a check executed via check_nt or check_nrpe?

I can put in a feature request. I guess the things that we're most concerned about re: "stability" is that our builtin defaults in the wizard *just work* and additionally that our documentation:

http://assets.nagios.com/downloads/nagi ... ios-XI.pdf

Still coincides with all the installation steps and such without too much confusion for rookie users of the product.

Can you guys (WillemDH, Box293) verify those to be the case? This is the kind of testing I was referring to us conducting before updating the package in the XI wizard.

With the "Windows Server Monitoring wizard", it uses check_nt command to execute check requests against windows servers and is used create all the service checks.

check_nt is no longer under development by the developer of NSClient++ and hence the module that worked in 0.3.9 continues to work exactly the same as in 0.4.1.105 and even the latest version 0.4.3.131.

The only thing requiring change in the documentation is screenshots to reflect the differences between the different versions.


When is comes to checks executed via check_nrpe, this is a whole different ballgame. The checks you execute via check_nrpe worked fine with 0.4.1.105. However 0.4.2.x and 0.4.3.x have undergone a major development re-write and they are still quite a few bugs being ironed out AS WELL as some commands and their arguments being completely rewritten. In saying that, I'm expecting that 0.4.4 or 0.4.5 will be pretty stable.

Checks offered by NSClient++ which are executed via check_nrpe are far superior to check_nt. There is a lot more flexibility and customisation available. Nagios XI does not have a wizard for NSClient++ nrpe checks however this would be a good wizard to create. It could easily account for the different versions and create the relevant commands based on the version that is detected.


Probably the most important thing to make a comment about is that at some stage in the future 0.3.9 will no longer work on newer versions of Microsoft Windows (for whatever compatibility reason). It is better to be ahead of the ball game instead of chasing your tail.

[WillemDH]

I follow Troy completely that check_nt should no longer be used. check_nrpe is much better, provides more options and is actively maintained and supported.

Migrating to 0.4.1.105 should not have to confuse 'rookie users', as, although I'm not using check_nt, I'm petty sure check_nt does still work with NSClient 0.4.1.105, as Troy said. Migrating check_nt to check_nrpe will be a little more work, as the wizard and documentation will need to get an update. Besides this, some components, as the Metrics component only work with check_nt...
But imho if Nagios did not already has plans to migrate from check_nt to check_nrpe, they should really start planning it.

To quote Michael Medin:

The main difference is that check_nt is limited to 11 command where as check_nrpe will given you access to many many more. You also have “more control” over the NRPE commands (as they allow arguments). There is also other technical aspects.

Nagios should not get behind with their NSClient agent version and used protocols. Although NCPA might become an alternative one day, some things are not (yet) possible with it. (like realtime eventlog monitoring, which really is a must have in any agent imho)

Grtz