Details
Description
From the log while doing test fest, the replicator is trying to save the checkpoint but the SG is not responding. The expectation is that replicator should be stopped / offline with timeout error. However it seems like CBL Replicator got change event from Lite Core Replicator with IDLE and timeout error.
2019-08-01 14:50:57.918253-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: <CBLWebSocket: 0x2828b9cc0>: HasBytesAvailable |
2019-08-01 14:50:57.918674-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: DoRead... |
2019-08-01 14:50:57.919134-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: DoRead read 82 bytes |
2019-08-01 14:50:57.919363-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: <<< received 82 bytes [now 82 pending] |
2019-08-01 14:50:57.920351-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: {BLIPIO#2968} Received frame: ERR #11 ----, length 78 |
2019-08-01 14:50:57.921135-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: {Connection#2966} Receiving ERR #11 |
2019-08-01 14:50:57.921350-0700 Todo[11966:4567949] CouchbaseLite Network Verbose: {Connection#2966} Finished receiving ERR #11 |
2019-08-01 14:50:57.922005-0700 Todo[11966:4567949] CouchbaseLite Replicator ERROR: {Repl#2967} Got error response: HTTP 503 'Database timeout error (gocb.ErrTimeout)' |
2019-08-01 14:50:57.922391-0700 Todo[11966:4567949] CouchbaseLite Replicator WARNING: {Repl#2967} Failed to save checkpoint! |
2019-08-01 14:50:57.922624-0700 Todo[11966:4567949] CouchbaseLite Replicator Verbose: {Repl#2967} progress +0/+0, 0 docs -- now 930 / 930, 4 docs |
2019-08-01 14:50:57.922830-0700 Todo[11966:4567949] CouchbaseLite Replicator Info: {Repl#2967} activityLevel=idle: connectionState=2 |
2019-08-01 14:50:57.923029-0700 Todo[11966:4567949] CouchbaseLite Replicator Info: {Repl#2967} now idle |
2019-08-01 14:50:57.923800-0700 Todo[11966:4567949] CouchbaseLite Replicator Info: CBLReplicator[<*> URL[ws://ec2-54-197-194-172.compute-1.amazonaws.com:4984/todo]] is idle, progress 930/930, error: Error Domain=CouchbaseLite Code=10503 "Database timeout error (gocb.ErrTimeout)" UserInfo={NSLocalizedDescription=Database timeout error (gocb.ErrTimeout)} |
2019-08-01 14:50:57.924402-0700 Todo[11966:4549482] [Todo] Replicator: IDLE 930/930, error: Error Domain=CouchbaseLite Code=10503 "Database timeout error (gocb.ErrTimeout)" UserInfo={NSLocalizedDescription=Database timeout error (gocb.ErrTimeout)} |
Attachments
Issue Links
- relates to
-
CBL-126 Litecore does not retry subChanges request that fails with temporary 503 error
-
- Closed
-
Activity
I think this is just a manifestation of CBL-126 because the library doesn't understand how to handle these kinds of messages inside of BLIP.
A question from Jim Borden, Does the iOS todo app have the functionality for me to reproduce this?
There is no specific functionality to reproduce this. The issue happened when SG had an issue (caused by Jay's large number of docs with attachment push). Instead of getting to stopped -> offline, the replicator went to IDLE with an error instead.
From the log, look like it happened after the replicator failed to save checkpoint.