Closed MatFl closed 5 years ago
When the server itself is reachable, but the service is stopped, the client receives connection refuse errors. E/CouchbaseLite/REPLICATOR: {Repl#1} Got LiteCore error: POSIX error 111 "Connection refused" The replicator then never enters STOPPED status, but keeps trying to reconnect without any backoff time.
Yes. This is intended behavior. It should, however, back off exponentially.
Is the application supposed to stop the replication itself when the error is received and the replicator is for example in OFFLINE state?
The application can do so if desired. The replication should stop after 2 retries.
... unless, of course, the replicator is configured continuous.
When the replicator is in this state for a while it keeps getting faster and faster with creating new connections.
Your log shows several different replicators running. I'm curious as to why you are running so many at once...
Is the application supposed to stop the replication itself when the error is received and the replicator is for example in OFFLINE state?
The application can do so if desired. The replication should stop after 2 retries.
... unless, of course, the replicator is configured continuous.
It is a continuous replication.
When the replicator is in this state for a while it keeps getting faster and faster with creating new connections.
Your log shows several different replicators running. I'm curious as to why you are running so many at once...
I only have one continuous push and pull replicator started. Whatever is happening here is done internally in the replicator.
I just noticed, that on a actual device this issue is not as critical. The incredible number of connections came from tests on an emulator. But I don't understand why this would make a difference.
However even on the phone, the replication sometimes keeps trying to create new connections, even when it is supposed to be stopped.
@MatFl Opened https://issues.couchbase.com/browse/CBL-131 to track this. Please follow it there.
Using CBL 2.5 we observed some crashes and strange behavior with continuous replication.
When the server itself is reachable, but the service is stopped, the client receives connection refuse errors.
E/CouchbaseLite/REPLICATOR: {Repl#1} Got LiteCore error: POSIX error 111 "Connection refused"
The replicator then never enters STOPPED status, but keeps trying to reconnect without any backoff time. Is the application supposed to stop the replication itself when the error is received and the replicator is for example in OFFLINE state?When the replicator is in this state for a while it keeps getting faster and faster with creating new connections.
The log is spammed with this:
As you can see it creates hundreds of retries in a very short amount of time.
When stopping the replication at this point, it does not really stop the process. It keeps trying to create new connections. After a while it crashes with this exception: