Is Elasticsearch a drop in replacement for Memcached?

Hi,
I've been trying to play with Elasticsearch as a drop in replacement for
Memcached. I've installed the memcached transport and it looks like it
started fine and is on the right port. However if I attempt to use the java
spy.memcahed client library I'm not able to cache things.

Instead I get errors like:

2014-10-01 08:22:31.579 INFO net.spy.memcached.MemcachedConnection: Added
{QA sa=localhost/127.0.0.1:11211, #Rops=0, #Wops=0, #iq=0, topRop=null,
topWop=null, toWrite=0, interested=0} to connect queue
2014-10-01 08:22:31.586 INFO net.spy.memcached.MemcachedConnection:
Connection state changed for sun.nio.ch.SelectionKeyImpl@76fb509a
2014-10-01 08:22:31.679 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=2, #Wops=0, #iq=0, topRop=Cmd: 2 Opaque: 2 Key: foo Cas: 0 Exp: 1000
Flags: 0 Data Length: 3, topWop=null, toWrite=0, interested=1}
java.io.IOException: Disconnected unexpected, will reconnect.

over and over.

On the Elasticsearch side I get an error of
[2014-10-01 08:22:31,675][ERROR][memcached.netty ] [Living Hulk]
Unsupported opcode [0x2], ignoring and closing connection

Here is my client configuration:
ConnectionFactoryBuilder factory = new ConnectionFactoryBuilder();
factory.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY);
SerializingTranscoder t = new SerializingTranscoder(1000000);
t.setCompressionThreshold(1024);
factory.setTranscoder(t);
factory.setOpTimeout(1000);
factory.setTimeoutExceptionThreshold(1998);
factory.setHashAlg(DefaultHashAlgorithm.KETAMA_HASH);
factory.setLocatorType(ConnectionFactoryBuilder.Locator.CONSISTENT);
factory.setFailureMode(FailureMode.Redistribute);
factory.setUseNagleAlgorithm(false);

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/c2b7c8d0-ff79-488c-b3c5-b8ca337bd955%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Hi,

Which versions of elasticsearch / elasticsearch-transport-memcached /
memcached are you using?

It looks like the memcached client is trying to execute a 0x02 ADD command
whereas the elasticsearch-transport-memcached only supports
get/set/delete/quit commands and disconnects for all others (see
documentation here:
https://github.com/elasticsearch/elasticsearch-transport-memcached)

You may try a "set" instead.

-- Tanguy

Le mercredi 1 octobre 2014 15:35:40 UTC+2, Ryan Bergman a écrit :

Hi,
I've been trying to play with Elasticsearch as a drop in replacement
for Memcached. I've installed the memcached transport and it looks like it
started fine and is on the right port. However if I attempt to use the java
spy.memcahed client library I'm not able to cache things.

Instead I get errors like:

2014-10-01 08:22:31.579 INFO net.spy.memcached.MemcachedConnection: Added
{QA sa=localhost/127.0.0.1:11211, #Rops=0, #Wops=0, #iq=0, topRop=null,
topWop=null, toWrite=0, interested=0} to connect queue
2014-10-01 08:22:31.586 INFO net.spy.memcached.MemcachedConnection:
Connection state changed for sun.nio.ch.SelectionKeyImpl@76fb509a
2014-10-01 08:22:31.679 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=2, #Wops=0, #iq=0, topRop=Cmd: 2 Opaque: 2 Key: foo Cas: 0 Exp: 1000
Flags: 0 Data Length: 3, topWop=null, toWrite=0, interested=1}
java.io.IOException: Disconnected unexpected, will reconnect.

over and over.

On the Elasticsearch side I get an error of
[2014-10-01 08:22:31,675][ERROR][memcached.netty ] [Living Hulk]
Unsupported opcode [0x2], ignoring and closing connection

Here is my client configuration:
ConnectionFactoryBuilder factory = new ConnectionFactoryBuilder();
factory.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY);
SerializingTranscoder t = new SerializingTranscoder(1000000);
t.setCompressionThreshold(1024);
factory.setTranscoder(t);
factory.setOpTimeout(1000);
factory.setTimeoutExceptionThreshold(1998);
factory.setHashAlg(DefaultHashAlgorithm.KETAMA_HASH);

factory.setLocatorType(ConnectionFactoryBuilder.Locator.CONSISTENT);
factory.setFailureMode(FailureMode.Redistribute);
factory.setUseNagleAlgorithm(false);

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/139a935a-fbc5-48fa-ab33-1a4c0cde4397%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

I'm using 1.3.3 with a transport of 2.3.0

I was attempting to use one of the standard common Memcache client
libraries that already exist. I was hoping that they would "just work"
without having to write my own client (or modifying one of them). It sounds
like that's not an option. Is there a plan for the transport to support ADD
in the future?

Ryan

On Thursday, October 2, 2014 2:27:58 AM UTC-5, Tanguy Leroux wrote:

Hi,

Which versions of elasticsearch / elasticsearch-transport-memcached /
memcached are you using?

It looks like the memcached client is trying to execute a 0x02 ADD command
whereas the elasticsearch-transport-memcached only supports
get/set/delete/quit commands and disconnects for all others (see
documentation here:
https://github.com/elasticsearch/elasticsearch-transport-memcached)

You may try a "set" instead.

-- Tanguy

Le mercredi 1 octobre 2014 15:35:40 UTC+2, Ryan Bergman a écrit :

Hi,
I've been trying to play with Elasticsearch as a drop in replacement
for Memcached. I've installed the memcached transport and it looks like it
started fine and is on the right port. However if I attempt to use the java
spy.memcahed client library I'm not able to cache things.

Instead I get errors like:

2014-10-01 08:22:31.579 INFO net.spy.memcached.MemcachedConnection:
Added {QA sa=localhost/127.0.0.1:11211, #Rops=0, #Wops=0, #iq=0,
topRop=null, topWop=null, toWrite=0, interested=0} to connect queue
2014-10-01 08:22:31.586 INFO net.spy.memcached.MemcachedConnection:
Connection state changed for sun.nio.ch.SelectionKeyImpl@76fb509a
2014-10-01 08:22:31.679 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=2, #Wops=0, #iq=0, topRop=Cmd: 2 Opaque: 2 Key: foo Cas: 0 Exp: 1000
Flags: 0 Data Length: 3, topWop=null, toWrite=0, interested=1}
java.io.IOException: Disconnected unexpected, will reconnect.

over and over.

On the Elasticsearch side I get an error of
[2014-10-01 08:22:31,675][ERROR][memcached.netty ] [Living Hulk]
Unsupported opcode [0x2], ignoring and closing connection

Here is my client configuration:
ConnectionFactoryBuilder factory = new ConnectionFactoryBuilder();
factory.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY);
SerializingTranscoder t = new SerializingTranscoder(1000000);
t.setCompressionThreshold(1024);
factory.setTranscoder(t);
factory.setOpTimeout(1000);
factory.setTimeoutExceptionThreshold(1998);
factory.setHashAlg(DefaultHashAlgorithm.KETAMA_HASH);

factory.setLocatorType(ConnectionFactoryBuilder.Locator.CONSISTENT);
factory.setFailureMode(FailureMode.Redistribute);
factory.setUseNagleAlgorithm(false);

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/f11aa1c9-17bc-4a30-8c97-bf09c003989e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

You don't need to write your own client, just to use SET commands instead
of ADD.

For example, with the spymemcached library you should do something like:

MemcachedClient c=new MemcachedClient(
new InetSocketAddress("hostname", portNum));

// Store a value (async) for one hour
c.set("someKey", 3600, someObject);
// Retrieve a value (synchronously).
Object myObject=c.get("someKey");

-- Tanguy

Le jeudi 2 octobre 2014 16:12:29 UTC+2, Ryan Bergman a écrit :

I'm using 1.3.3 with a transport of 2.3.0

I was attempting to use one of the standard common Memcache client
libraries that already exist. I was hoping that they would "just work"
without having to write my own client (or modifying one of them). It sounds
like that's not an option. Is there a plan for the transport to support ADD
in the future?

Ryan

On Thursday, October 2, 2014 2:27:58 AM UTC-5, Tanguy Leroux wrote:

Hi,

Which versions of elasticsearch / elasticsearch-transport-memcached /
memcached are you using?

It looks like the memcached client is trying to execute a 0x02 ADD
command whereas the elasticsearch-transport-memcached only supports
get/set/delete/quit commands and disconnects for all others (see
documentation here:
https://github.com/elasticsearch/elasticsearch-transport-memcached)

You may try a "set" instead.

-- Tanguy

Le mercredi 1 octobre 2014 15:35:40 UTC+2, Ryan Bergman a écrit :

Hi,
I've been trying to play with Elasticsearch as a drop in replacement
for Memcached. I've installed the memcached transport and it looks like it
started fine and is on the right port. However if I attempt to use the java
spy.memcahed client library I'm not able to cache things.

Instead I get errors like:

2014-10-01 08:22:31.579 INFO net.spy.memcached.MemcachedConnection:
Added {QA sa=localhost/127.0.0.1:11211, #Rops=0, #Wops=0, #iq=0,
topRop=null, topWop=null, toWrite=0, interested=0} to connect queue
2014-10-01 08:22:31.586 INFO net.spy.memcached.MemcachedConnection:
Connection state changed for sun.nio.ch.SelectionKeyImpl@76fb509a
2014-10-01 08:22:31.679 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=2, #Wops=0, #iq=0, topRop=Cmd: 2 Opaque: 2 Key: foo Cas: 0 Exp: 1000
Flags: 0 Data Length: 3, topWop=null, toWrite=0, interested=1}
java.io.IOException: Disconnected unexpected, will reconnect.

over and over.

On the Elasticsearch side I get an error of
[2014-10-01 08:22:31,675][ERROR][memcached.netty ] [Living
Hulk] Unsupported opcode [0x2], ignoring and closing connection

Here is my client configuration:
ConnectionFactoryBuilder factory = new ConnectionFactoryBuilder();
factory.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY);
SerializingTranscoder t = new SerializingTranscoder(1000000);
t.setCompressionThreshold(1024);
factory.setTranscoder(t);
factory.setOpTimeout(1000);
factory.setTimeoutExceptionThreshold(1998);
factory.setHashAlg(DefaultHashAlgorithm.KETAMA_HASH);

factory.setLocatorType(ConnectionFactoryBuilder.Locator.CONSISTENT);
factory.setFailureMode(FailureMode.Redistribute);
factory.setUseNagleAlgorithm(false);

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/106b7c04-04db-499a-9204-6d57d72a7278%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Sure, but I'll have to go through all the existing prod code that uses the
client and change it to set. It would be nicer in the future if add was
supported. I noticed there is a github issue for
that https://github.com/elasticsearch/elasticsearch-transport-memcached/issues/7

Also, in my test scenario I'm now having other issues with the spy.memcache
client. It's a little obscure and happens on a get:
2014-10-02 09:53:08.931 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=1, #Wops=0, #iq=0, topRop=Cmd: 0 Opaque: 3 Key: foo, topWop=null,
toWrite=0, interested=1}
java.lang.ArrayIndexOutOfBoundsException: 0
at
net.spy.memcached.protocol.binary.OperationImpl.decodeInt(OperationImpl.java:257)
at
net.spy.memcached.protocol.binary.GetOperationImpl.decodePayload(GetOperationImpl.java:51)
at
net.spy.memcached.protocol.binary.OperationImpl.finishedPayload(OperationImpl.java:178)
at
net.spy.memcached.protocol.binary.OperationImpl.readFromBuffer(OperationImpl.java:162)
at
net.spy.memcached.MemcachedConnection.handleReads(MemcachedConnection.java:463)
at
net.spy.memcached.MemcachedConnection.handleIO(MemcachedConnection.java:380)
at
net.spy.memcached.MemcachedConnection.handleIO(MemcachedConnection.java:242)
at net.spy.memcached.MemcachedConnection.run(MemcachedConnection.java:836)

Looks like it's happening in a method that is reading the binary data:

static int decodeInt(byte[] data, int i) {
return (data[i] & 0xff) << 24
| (data[i + 1] & 0xff) << 16
| (data[i + 2] & 0xff) << 8
| (data[i + 3] & 0xff);
}

I'll try and dig in a little more and also maybe try another client like
the Whalin client.

Ryan

On Thursday, October 2, 2014 9:36:02 AM UTC-5, Tanguy Leroux wrote:

You don't need to write your own client, just to use SET commands instead
of ADD.

For example, with the spymemcached library you should do something like:

MemcachedClient c=new MemcachedClient(
new InetSocketAddress("hostname", portNum));

// Store a value (async) for one hour
c.set("someKey", 3600, someObject);
// Retrieve a value (synchronously).
Object myObject=c.get("someKey");

-- Tanguy

Le jeudi 2 octobre 2014 16:12:29 UTC+2, Ryan Bergman a écrit :

I'm using 1.3.3 with a transport of 2.3.0

I was attempting to use one of the standard common Memcache client
libraries that already exist. I was hoping that they would "just work"
without having to write my own client (or modifying one of them). It sounds
like that's not an option. Is there a plan for the transport to support ADD
in the future?

Ryan

On Thursday, October 2, 2014 2:27:58 AM UTC-5, Tanguy Leroux wrote:

Hi,

Which versions of elasticsearch / elasticsearch-transport-memcached /
memcached are you using?

It looks like the memcached client is trying to execute a 0x02 ADD
command whereas the elasticsearch-transport-memcached only supports
get/set/delete/quit commands and disconnects for all others (see
documentation here:
https://github.com/elasticsearch/elasticsearch-transport-memcached)

You may try a "set" instead.

-- Tanguy

Le mercredi 1 octobre 2014 15:35:40 UTC+2, Ryan Bergman a écrit :

Hi,
I've been trying to play with Elasticsearch as a drop in replacement
for Memcached. I've installed the memcached transport and it looks like it
started fine and is on the right port. However if I attempt to use the java
spy.memcahed client library I'm not able to cache things.

Instead I get errors like:

2014-10-01 08:22:31.579 INFO net.spy.memcached.MemcachedConnection:
Added {QA sa=localhost/127.0.0.1:11211, #Rops=0, #Wops=0, #iq=0,
topRop=null, topWop=null, toWrite=0, interested=0} to connect queue
2014-10-01 08:22:31.586 INFO net.spy.memcached.MemcachedConnection:
Connection state changed for sun.nio.ch.SelectionKeyImpl@76fb509a
2014-10-01 08:22:31.679 INFO net.spy.memcached.MemcachedConnection:
Reconnecting due to exception on {QA sa=localhost/127.0.0.1:11211,
#Rops=2, #Wops=0, #iq=0, topRop=Cmd: 2 Opaque: 2 Key: foo Cas: 0 Exp: 1000
Flags: 0 Data Length: 3, topWop=null, toWrite=0, interested=1}
java.io.IOException: Disconnected unexpected, will reconnect.

over and over.

On the Elasticsearch side I get an error of
[2014-10-01 08:22:31,675][ERROR][memcached.netty ] [Living
Hulk] Unsupported opcode [0x2], ignoring and closing connection

Here is my client configuration:
ConnectionFactoryBuilder factory = new ConnectionFactoryBuilder();
factory.setProtocol(ConnectionFactoryBuilder.Protocol.BINARY);
SerializingTranscoder t = new SerializingTranscoder(1000000);
t.setCompressionThreshold(1024);
factory.setTranscoder(t);
factory.setOpTimeout(1000);
factory.setTimeoutExceptionThreshold(1998);
factory.setHashAlg(DefaultHashAlgorithm.KETAMA_HASH);

factory.setLocatorType(ConnectionFactoryBuilder.Locator.CONSISTENT);
factory.setFailureMode(FailureMode.Redistribute);
factory.setUseNagleAlgorithm(false);

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/ba738dbb-b325-4df1-8f75-47faeafa38a2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.