Question

I'm studying the following scenario: while get/set operations to Couchbase I shutting down node(power off on virtual machine). After that, I power on the machine and waiting for Couchbase node recovery. When node's status changing to "healthy" I expect that client reconnect and get/set operations continues. But sometimes reconnection of client occurs immediately, sometime doesn't occur within a few minutes.

So my question is: Are there some configuration on server side, or on client side that guarantee a wholly reconnection of client?

I use JavaSDK.

A small addition: Couchbase client is based on spymemcached client. If someone knows any hints with memcached, that could solve problem, I'll be very glad to see them.

Another addition:

Client stops trying to establish connection after this exception:

Exception in thread "Thread-122" java.lang.IllegalStateException: Got empty SASL auth mech list. 11:59:25,731 ERROR [stderr] (Thread-122) at net.spy.memcached.auth.AuthThread.listSupportedSASLMechanisms(AuthThread.java:99) 11:59:25,731 ERROR [stderr] (Thread-122) at net.spy.memcached.auth.AuthThread.run(AuthThread.java:112)

But I can't understand, why this exception happens so irregularly.

Was it helpful?

Solution

It's a small bug with response of CouchbaseServer. It's my discussion of this problem with developer of spymemcached client. Problem was solved when I hard-coded some string at this client.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top