Hello Markus,
The first question:
In LV API, there's a VI "Delete_Process.vi". This VI just removes the Process from Shared Varibale Engine. If i do open the MAX Historical Dataviewer i still can see the entry for the deleted process, but the "Get Process List.vi" doesn't show the deleted process.
Now what's happening with this process? Is there a something like a Garbage Collector?
How can i remove the data from this deleted process from disc?
Which kind of view creates MAX Historical Data Viewer?
You are right - the process is only removed from the list, as the data traces still exist. You might want to delete the traces first, e.g. using the 'Delete Traces.vi' from the [DSC Module]»[Historical]»[Database Management] palette.
The second question:
I create a process every night at 0:00. Each process consists of 150 Traces. Lifespan of the process is 12 days.
On the same system there's a Lookout installation.
The problem now is, that the Shared Varibale Browser in Lookout now shows all of these Traces for each process.
That's very bad, because there are now 1800 more Varibales.
Is there a possibility to write into Traces, that don't appear in the Shared Variable Browser?
Are there any attributes that can be set?
I would like to use the Citadel Database, but without the Shared Varibale Engine. Is there a "Lowlevel Access" possible?
You use the 'Open Trace' from the [DSC Module]»[Historical]»[Database Writing] palette to create a new trace for a specified process and write the data directly to the database. This should omit the display in the shared variable browser, although I have to admit that I could not verify this yet.
The third question:
Lifespan for the process is set to 12 days. But the entrys in MAX Historical Dataviewer doesn't disappear.
OK, I know that Citadel removes these process maybe same days later...
Is there somewhere an option that defines a max. Size for the Database if i only had a small disk capacity?
This matter is a bit more complex then first thought. The lifespan only defines the minimum lifetime of the process's data, and additionally, is only relevant within a process. So, if you create new processes every night, your old data is "saved" forever, as only the original process can discard traces after they reach their lifespan.
The size of the database cannot be limited automatically as far as I know, but you could use archiving to move older processes out of the current database. If not needed, you could delete the archived databases every n days programmatically, so your working data stays small.
One more thing to think about: why do you create new processes every night? If your goal is only the identification of separate test runs, you might use dataset marking throughout a continuously running process. This might eliminate all your problems at once: As the process count stays low, the identification of your data points in lookout will be a lot easier, and the lifespan of your traces will be used to keep the database smaller...
There is an example project included in the LabVIEW DSC Module, but at the moment it only works in an english language environment, as some named references have been hard-coded in it. You might still have a look at it to see how it is done...
Best regards,
Sebastian
