toring piggyback data for: 'NVIDIA-DLS-3'
Trying to acquire lock on /omd/sites/cmk/tmp/check_mk/piggyback/NVIDIA-DLS-3/NTX
Got lock on /omd/sites/cmk/tmp/check_mk/piggyback/NVIDIA-DLS-3/NTX
Releasing lock on /omd/sites/cmk/tmp/check_mk/piggyback/NVIDIA-DLS-3/NTX
Released lock on /omd/sites/cmk/tmp/check_mk/piggyback/NVIDIA-DLS-3/NTX
Storing piggyback data for: 'QUIRM_-_WSUS'
Trying to acquire lock on /omd/sites/cmk/tmp/check_mk/piggyback/QUIRM_-_WSUS/NTX
Got lock on /omd/sites/cmk/tmp/check_mk/piggyback/QUIRM_-_WSUS/NTX
Releasing lock on /omd/sites/cmk/tmp/check_mk/piggyback/QUIRM_-_WSUS/NTX
Released lock on /omd/sites/cmk/tmp/check_mk/piggyback/QUIRM_-_WSUS/NTX
Storing piggyback data for: 'EX2016'
Trying to acquire lock on /omd/sites/cmk/tmp/check_mk/piggyback/EX2016/NTX
Got lock on /omd/sites/cmk/tmp/check_mk/piggyback/EX2016/NTX
Releasing lock on /omd/sites/cmk/tmp/check_mk/piggyback/EX2016/NTX
Released lock on /omd/sites/cmk/tmp/check_mk/piggyback/EX2016/NTX
Also inside the tmp/check_mk/piggyback no data can be found.
Header inside the agent output looks like this.
<<<<Quirm - WSUS>>>>
<<<prism_vm:sep(0)>>>
The error also happens if you have no piggyback hostname translation rule active.
In my example was the rule set to uppercase for all names.
We are validating the host names now, and ignoring all invalid names. We had “names” like ".." and other nonsense in the past, which are not very good file names .
But it should be ok if your translation changes it into a valid name, I think. Is that the case?
Looks like this was removed by accident. However, I think it actually is better. We should not on one hand allow the configuration of translations, and then on the other hand do some translations automagically.
So I think probably I’ll just add an incompatible Werk.
What about the translation rule? It seems somewhat insufficient to me – but these things are sometimes hard to judge from a developers perspective. Is it practical to configure needed translations? Or is something missing?
I’ll definitively talk to product management and the UX team about this.
We can extend the ruleset to give users more options – from a technical (dev) perspective, that is easy. But I want to be careful with that, and only add a small number of options that give the most benefit.
What are common requirements? Can you give more concrete examples?
The normal situation is, that no one knows what names you get as piggyback names.
Inside VMWare nearly everything is allowed for the naming, i think that was also the reason for the extra options inside the special agent.
That is the same for backup inside Veeam or like my example Nutanix.
From my point of view it would be better to have the same processing for the names for all sources of piggyback data. That also means the special agent reports the names as they are inside the data source.
Why not have the same behavior as before or inside some classic Monitoring Tools?
Replace invalid characters with underscore or remove invalid characters completely.
But not ignore the data.