Open GoogleCodeExporter opened 9 years ago
I have this same problem with 2.8.4 and 2.8.9. Using wireshark, I was able to
determine that spymemcached is sending a noop operation to the server without
first sending the authentication. I suspect this is happening because of a race
condition between the call to connect and the registration of the OP_CONNECT
selection key.
Original comment by cgri...@gmail.com
on 3 Apr 2013 at 4:04
Hi,
the noop has been implemented to make sure that we can detect "zombie" memcache
processes, for example those that are responding on the socket but are actually
dead.
We ran the tests against both vanilla memcached and also couchbase buckets.
Is it possible that you guys provide mor info on your environment so we can
debug the issue?
THanks!
Original comment by michael....@gmail.com
on 3 Apr 2013 at 6:52
Thanks for your reply,
I reproduced the problem on my MBP Retina 2012 running Mountain Lion and
Couchbase 2.0 installed via Homebrew.
Please let me know if you need anything else!
Original comment by yku...@gmail.com
on 3 Apr 2013 at 1:08
I recreate this issue 100% of the time on my MacBook Pro (OS 10.8.3) and on
Heroku connecting to either of free accounts provided by Heroku (via NorthScale
aka Couchbase) or Memcachier.
Mac Java:
java version "1.6.0_43"
Java(TM) SE Runtime Environment (build 1.6.0_43-b01-447-11M4203)
Java HotSpot(TM) 64-Bit Server VM (build 20.14-b01-447, mixed mode)
Heroku Java:
java version "1.6.0_27"
OpenJDK Runtime Environment (IcedTea6 1.12.4) (Ubuntu build 1.6.0_27-b27)
OpenJDK 64-Bit Server VM (build 20.0-b12, mixed mode)
Original comment by cgri...@gmail.com
on 3 Apr 2013 at 3:51
I have the same problem reported by original ticket opener.
Server framework is PlayFramework 2.1 connecting to memcached server using the
https://github.com/mumoshu/play2-memcached module. Everything worked fine
until play2-memcached module upgraded to spymemcached 2.8.10. Development
environment is windows 7, Java 7_15, Couchbase 2.0. Authentication issue
happens on both their couchbase bucket and memcached bucket.
Let me know if there is anything else I can provide that could help.
Original comment by enalm...@gmail.com
on 3 Apr 2013 at 7:27
This issue has also been reported at
https://groups.google.com/forum/?fromgroups=#!topic/clojure-memcached/_wBnwRXwb3
M
Original comment by cgri...@gmail.com
on 3 Apr 2013 at 7:35
To those also blocked by this issue, I was able to switch to xmemcached using
AuthInfo.plain (AuthInfo.typical did not work for my servers) and get my app
working.
Original comment by cgri...@gmail.com
on 18 Apr 2013 at 4:20
I'll look into this today and report my findings. We added some stuff in 2.8.10
to make sure connections are established properly which may be part of the
problem reported here.
Original comment by michael....@gmail.com
on 19 Apr 2013 at 6:36
Okay, so please use port 11210 if you are doing SASL auth against a memcached
bucket on a couchbase system!
AuthDescriptor authDescriptor = new AuthDescriptor(new String[]{"PLAIN"},
new PlainCallbackHandler("cache", "cache"));
ConnectionFactory fact = new ConnectionFactoryBuilder()
.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY)
.setAuthDescriptor(authDescriptor)
.build();
MemcachedClient client = new MemcachedClient(fact, Arrays.asList(new InetSocketAddress("127.0.0.1", 11210)));
client.set("foo", 0, "value");
System.out.println(client.get("foo"));
client.shutdown();
Code like this should work if you have a memcached bucket called "cache" with a
"cache" password. Let me know if this works for you!
11211 is moxi for the default bucket, you don't need to have SASL in place
there.
Let me know if this works for you guys, watch out the memcache ports against a
couchbase server!
Original comment by michael....@gmail.com
on 19 Apr 2013 at 12:01
Original issue reported on code.google.com by
yku...@gmail.com
on 31 Mar 2013 at 3:55