It seems that the AphlictMaster will happily "connect" to the Aphlict server, even if the Aphlict server (that is, the NodeJS server) is not running. Possibly the AphlictMaster should raise some sort of error in this case (notifications are enabled but don't seem to be working) or, at least, the AphlictMaster should just die quietly.
Description
Description
Revisions and Commits
Revisions and Commits
| Status | Assigned | Task | ||
|---|---|---|---|---|
| Resolved | joshuaspence | T5365 Aphlict client should reconnect after a failure | ||
| Resolved | epriestley | T5281 Aphlict master doesn't handle connection failures |
Event Timeline
Comment Actions
This might be tricky to handle correctly.
It seems that, if the AphlictMaster fails to connect to the Aphlict server after 20 seconds (the default socket timeout), a SecurityErrorEvent will be triggered. I'm not quite sure how we can differentiate between a SecurityErrorEvent triggered by a connection failure versus a SecurityErrorEvent triggered by a legitimate security error.
Source: http://helpx.adobe.com/flash/kb/flash-player-10-socket-connection.html
Comment Actions
Also, this explains why users have been reporting that Aphlict was generating an "Error #2048" error. This was largely fixed by D9418, but the root cause was that the Aphlict client was being created (even though notifications were disabled) and failing to connect to the (non-running) Aphlict server.
Comment Actions
(I think the underlying cause of the security error is that the Aphlict server isn't running on port 843, so the client isn't able to read the policy file)
Comment Actions
I'm going to close this since we have a way to do something mostly reasonable-ish now (i.e., not scary and confusing) and do something reasonable-ish for the most common issue (flash policy server issues). The actual behavior in response to various errors can be improved, but T5365 can cover that.