Symptoms
An OA managed host is unreachable.
Many tasks, related to the host, are failing with the error:
Cannot establish connection to the host #NNN
There is an existing record in unreachable_hosts
table of OA DB for the host with a large amount of failures:
plesk=> select * from unreachable_hosts ;
host_id | checked_at | c_failures | last_response_delay
---------+------------------------+------------+---------------------
35 | 2016-05-28 01:37:40+08 | 116 | 1
OA Agent was down on the host, but got started back, which did not resolve the issue with tasks.
Cause
If slave node was unreachable for a long time, it does not become available right away after restart, because of a hardcoded timeout of unreachable hosts checking. The timeout becomes longer in proportion to the number of failed attempts to connect.
Resolution
Wait more to have the OA Management Node reconnect to the slave host and recognize that it is available.
In case of time constraints, contact Odin Technical Support to restore the host connectivity instantly.