Hello!
I have a problem with remote probes sporadically losing connection to the PRTG Core server. Their logs show something like this:
27.09.2012 10:55:23 Login OK: Willkommen bei PRTG
30.09.2012 13:08:05 Disconnected from 10.2.9.34:23560
30.09.2012 13:08:13 Socket Error # 10054 Connection reset by peer. (will keep trying)
30.09.2012 13:08:23 Create new Connection
30.09.2012 13:08:48 Create new Connection
30.09.2012 13:09:12 Create new Connection
30.09.2012 13:09:37 Create new Connection
30.09.2012 13:10:01 Create new Connection
30.09.2012 13:10:25 Create new Connection
30.09.2012 13:10:50 Create new Connection
30.09.2012 13:11:14 Create new Connection
30.09.2012 13:11:38 Create new Connection
30.09.2012 13:12:03 Create new Connection
30.09.2012 13:12:16 10 connection attempts failed with error Socket Error # 10054 Connection reset by peer. (will keep trying)
Now the interesting thing - when I log on to the PRTG core server with RDP (or log off and then log on again if a session was already open) the remote probes log back in. This is reproducible, creating a new session on the PRTG server always solves the disconnect. Of course this can't be a permanent solution.
Do you know of any similar cases and a way this might be fixed?
Regards, Christian
Article Comments
Hello,
the error message "Socket Error # 10054 Connection reset by peer" hints on a firewall or something similar blocking the connection, which then may abandon the rule (which is blocking) as soon as an RDP-Connection to this Probe Host is opened. Sounds very strange, and we have to say that so far we have not encountered a similar issue to be honest.
best regards.
Oct, 2012 - Permalink
I can confirm this is happening to me as well. No error message other than the disconnect.
5/28/2013 3:53:57 PM Probe Health Up 100 % 5/28/2013 3:53:23 PM Probe Health Down Disconnected
I would log into to the server that is holding the remote probe and after a certain amount of time, the probe will disconnect, and when I log into the server to check it, the server reconnects and the alarms go away.
May, 2013 - Permalink
Are there firewalls between the core server and that probe?
In that case - ask you nertwork admins if your firewalls do something like DPI-SSL (as it is called on Dell Sonicwalls). That was the problem for us - it does a kind of "man in the middle" on SSL connections to inspect encrypted traffic. And somehow DPI-SSL and PRTG don't seem to go together very well - we disabled DPI-SSL for the probe-port and had no problems ever since.
Regards, Christian
May, 2013 - Permalink
Please open a Support Ticket or send an email to support@paessler.com as regards this issue.
Oct, 2012 - Permalink