Working with management API is affecting Management.

[Oscar_David_Gom]

Hi,

I'm working with a KVM machine with RHEL 7 and Gaia R80.20 GA, I'm working on a migration of a management R80.10 whos database gets corrupted in some strange way (Everything is working ok and suddenly SmartConsole closes and I can't access to it because an internal error, when I look into $CPDIR/registry I get a lot of corrupted files, I used sk 105510 in order to gain acces again an it works, but it continues happen constantly.

The migration that i'm doing is exporting a .csv with the objects from r80.10 mgmt and then importing to the new r80.20 mgmt with mgmt API (This is because I dont want to migrate any error of the corrupted database from r80.10 mgmt), yesterday i was in this process and the r80.20 mgmt started to behave like the r80.10.

I want to know if there is a posibility that multiple actions of importing with MGMT API can affect in some way the mangamenet database, maybe host objects with same IP address? or objects which name has a format like "network_1" (I know sk113078 says that _ can cause an error but in other migrations I've never have had problems) 

Thanks

[PhoneBoy]

In general, if you're importing things thru CSV files, you want to limit them to no more than 500 lines.
Likewise, when you're working with the API in bulk, you want to make sure you issue periodic publish commands.
You also want to make sure you've allocated enough RAM/CPU to your management VMs.
Even though the "minimum" in the release notes says 6GB RAM and 2 cores, I highly recommend at least 16GB RAM and 4 cores as the minimum.