BLIP+WebSocket connection error: unexpected EOF

#1

Hi Guys,

We are getting EOF , which is stopping sync , and changes stops coming.

When this happens I see #4116: --> BLIP+WebSocket connection error: unexpected EOF in sync gateway logs.

Can you help what it is .

CC: @priya.rajagopal, @househippo
Regards
Pankaj Sharma

#2

can we get full logs?

#3

If possible, please share both client and SG log. Thanks.

#4

After removal of conflict documents this issue got resolved. If this comes again then only I can provide client logs. The sync gateway only had this much log, apart from this there were logs of get only.

#5

This has come back again, on Device the error is shown as : Websocket connection closed by peer and Same EOF file exception I can see on sync gateway.

It may be possibility that EOF is not the reason, But I am not sure.

Regards
Pankaj Sharma

#6

Just for your information we are using Views on couchbase side, And we had faced issues like that document is not in data service but the key is still present in channel view. In this case the document will not be present.

#7

Above is the screen shot of error on device, I will share the text version soon.

Also got following error too BLIP+WebSocket connection error: flate: corrupt input before offset 4094

@jens Can you help ?

#8

Did you submit client and server logs as Pasin requested last week?

#9

I have posted whatever logs I was getting.

#10

You still have not provided these.

#11

Following is the error log :

03-12 16:28:45.021 5732-5748/com.couchbase.todo I/LiteCoreJNI: [NATIVE] C4Replicator.statusChangedCallback() repl -> 0x0xa408bf80 status -> 4
03-12 16:28:45.022 5732-5750/com.couchbase.todo I/LiteCoreJNI: [NATIVE] C4Replicator.getResponseHeaders() repl -> 0xa408bf80
03-12 16:28:45.769 5732-5749/com.couchbase.todo I/LiteCoreJNI: [NATIVE] C4Replicator.statusChangedCallback() repl -> 0x0xa408bf80 status -> 3
03-12 16:33:46.281 5732-5752/com.couchbase.todo W/WS: WebSocketListener.onFailure() response -> null
java.io.EOFException
at okio.RealBufferedSource.require(RealBufferedSource.java:60)
at okio.RealBufferedSource.readByte(RealBufferedSource.java:73)
at okhttp3.internal.ws.WebSocketReader.readHeader(WebSocketReader.java:113)
at okhttp3.internal.ws.WebSocketReader.processNextFrame(WebSocketReader.java:97)
at okhttp3.internal.ws.RealWebSocket.loopReader(RealWebSocket.java:265)
at okhttp3.internal.ws.RealWebSocket$2.onResponse(RealWebSocket.java:204)
at okhttp3.RealCall$AsyncCall.execute(RealCall.java:153)
at okhttp3.internal.NamedRunnable.run(NamedRunnable.java:32)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636)
at java.lang.Thread.run(Thread.java:764)
03-12 16:33:46.348 5732-5748/com.couchbase.todo E/LiteCore [Sync]: {Repl#1}==> N8litecore4repl10ReplicatorE /data/user/0/com.couchbase.todo/files/dxpprodappdocstore.cblite2/ ->wss:dev-sagar.dxp-decurtis.com/syncgateway/prodappdocstore/_blipsync
{Repl#1} Got LiteCore error: Connection reset by peer (2/104)

#12

Please find attached logs in zip file. Whats happening is it connects then EOF exception comes at sync gateway and connection gets closed. then retry happens again.

clientlogs.log.zip (20.3 KB)

@pasin will this help.

#13

following are the sync gateway logs : sync_error.log.zip (9.0 KB)

#14

As a rule of thumb, if one side of the connection fails with an unexpected EOF, then the problem is some kind of fatal error on the other side that caused it to drop the connection immediately. Often this is a crash / panic or unexpected exception. So the log with the EOF error is usually not interesting; it’s the other side you want to look at.

#15

This log is from the sync gateway which can be helpful.

#16

@pankaj.sharma This looks like a problem that could be better managed via Couchbase Support - if you have an SLA, I’d recommend directing the question there.

From the details provided it’s unclear whether this is an EOF that would be expected to be seen if the connection between Sync Gateway and the client was dropped mid-replication, or an EOF being encountered for an otherwise healthy connection.

#17

The problem is I have flushed the buckets as of now which seems to clear the issue.

closed #18
unlisted #19