Replies: 3 comments 2 replies
-
Hi @dtuecks thanks for the report, let's have a closer look... actually the first thing I want to check... are you using device_no or devices_no in your config? the setting name is the second form (plural): devices_no and devices_only regards |
Beta Was this translation helpful? Give feedback.
-
Hi @dtuecks OK let's move on to debugging then :) So the way devices_* works is simply to copy the content into the four other settings (discover, macsuck, arpnip, netbios) when any Netdisco process starts. Are you able to give me the output of the following:
And then also when a backend worker Manager process starts, it should refresh the skiplist. So it is normal that the skiplist has things in, but they should be there from the start. If you look again in the database, are they assigned to the correct "backend" (this should be the FQDN of the host system running the workers) ? thanks |
Beta Was this translation helpful? Give feedback.
-
Just to add a little extra info on the original discussion question - no, the settings should be local to the worker and not global. Each of the processes reads its own local configuration file. The database skiplist table has a "backend" field which allows each backend to only get the skipped devices for its own environment. When the backend starts up, the Manager process rebuilds the skiplist. The workers should then only ever select jobs which are allowed for that backend. However, there are several cases where a backend worker can select a job and then discover it's not permitted to run. It should not attempt to make connection ever, though, and it may update the skiplist. |
Beta Was this translation helpful? Give feedback.
-
Hello,
we have a netdisco installation with several workers (6 in total). One of these workers is placed in a DMZ so that it has access to some devices that the others don't and vice versa.
The Idea was to use a combination of device_only / device_no ACLs. But since I've seen that "wrong" devices are fetched by the DMZ worker (device_no: 10.0.0.0/8 and it fetched 10.0.0.1) and then appear in the device_skip table (because DMZ worker can't reach 10.0.0.1), I suspect that this is a global setting, even if configured at just one worker.
Am I right? If so, any ideas how to 'fix' this?
Regards
Daniel
Beta Was this translation helpful? Give feedback.
All reactions