Jenkins build is back to normal : Cassandra-quick #59
See https://builds.apache.org/job/Cassandra-quick/59/changes
Re: [VOTE] Release Apache Cassandra 1.0.0 (Final) - Strike 2
The artifacts for this vote happen to have https://issues.apache.org/jira/browse/CASSANDRA-3343. This is a purely cosmetic issue, and clearly I don't want to reroll the vote for this, which would delay the release. But I also think that fixing this issue has zero chance of screwing anything up, so while it feels stupid to delay the release for that, it also feels stupid to me to annoy people with it. So I have created new artifacts, based on the previous ones but with the infamous 'echo $CLASSPATH' removed from cassandra.in.sh. And what I'm proposing is to simply switch to those last artifacts for the release instead of the initial ones, but without restarting a vote just for that. The new artifacts are: SVN: https://svn.apache.org/repos/asf/cassandra/branches/cassandra-1.0.0@r1181741 Artifacts: https://repository.apache.org/content/repositories/orgapachecassandra-053/org/apache/cassandra/apache-cassandra/1.0.0/ Staging repository: https://repository.apache.org/content/repositories/orgapachecassandra-053/ I've also put them and the corresponding debian package at http://people.apache.org/~slebresne/with_3343/ We still have 2 days to go on this vote, so this is not a sneaky move either. However, I've kept the initial artifacts around for now, so if someone disagree strongly, we'll stick to the initial artifacts. -- Sylvain On Tue, Oct 11, 2011 at 7:30 AM, Todd Burruss bburr...@expedia.com wrote: This is a bit preliminary, but wanted to get this to you guys knowing the vote is in progress using these artifacts I am seeing the following exception on restart. (also with RC1 and RC2.) the only interesting tidbit is that it seems to only happen when writing via direct calls to StorageProxy. I can't quite compare apples to apples, but if I only write via Thrift interface, replay of the commitlog works. Write via StorageProxy (but a different CF than via Thrift) the replay throws exception. INFO 22:06:50,671 Replaying /var/lib/cassandra/commitlog/CommitLog-1318309128835.log, /var/lib/cassandra/commitlog/CommitLog-1318309569302.log INFO 22:06:50,700 Finished reading /var/lib/cassandra/commitlog/CommitLog-1318309128835.log ERROR 22:06:50,702 Exception encountered during startup org.apache.cassandra.db.ColumnSerializer$CorruptColumnException: invalid column name length 0 at org.apache.cassandra.db.ColumnSerializer.deserialize(ColumnSerializer.java: 89) at org.apache.cassandra.db.ColumnSerializer.deserialize(ColumnSerializer.java: 36) at org.apache.cassandra.db.ColumnFamilySerializer.deserializeColumns(ColumnFam ilySerializer.java:143) at org.apache.cassandra.db.ColumnFamilySerializer.deserializeColumns(ColumnFam ilySerializer.java:135) at org.apache.cassandra.db.ColumnFamilySerializer.deserialize(ColumnFamilySeri alizer.java:128) at org.apache.cassandra.db.RowMutation$RowMutationSerializer.deserialize(RowMu tation.java:404) at org.apache.cassandra.db.commitlog.CommitLog.recover(CommitLog.java:284) at org.apache.cassandra.db.commitlog.CommitLog.recover(CommitLog.java:171) at org.apache.cassandra.service.AbstractCassandraDaemon.setup(AbstractCassandr aDaemon.java:196) at org.apache.cassandra.service.AbstractCassandraDaemon.activate(AbstractCassa ndraDaemon.java:337) at org.apache.cassandra.thrift.CassandraDaemon.main(CassandraDaemon.java:106) org.apache.cassandra.db.ColumnSerializer$CorruptColumnException: invalid column name length 0 at org.apache.cassandra.db.ColumnSerializer.deserialize(ColumnSerializer.java: 89) at org.apache.cassandra.db.ColumnSerializer.deserialize(ColumnSerializer.java: 36) at org.apache.cassandra.db.ColumnFamilySerializer.deserializeColumns(ColumnFam ilySerializer.java:143) at org.apache.cassandra.db.ColumnFamilySerializer.deserializeColumns(ColumnFam ilySerializer.java:135) at org.apache.cassandra.db.ColumnFamilySerializer.deserialize(ColumnFamilySeri alizer.java:128) at org.apache.cassandra.db.RowMutation$RowMutationSerializer.deserialize(RowMu tation.java:404) at org.apache.cassandra.db.commitlog.CommitLog.recover(CommitLog.java:284) at org.apache.cassandra.db.commitlog.CommitLog.recover(CommitLog.java:171) at org.apache.cassandra.service.AbstractCassandraDaemon.setup(AbstractCassandr aDaemon.java:196) at org.apache.cassandra.service.AbstractCassandraDaemon.activate(AbstractCassa ndraDaemon.java:337) at org.apache.cassandra.thrift.CassandraDaemon.main(CassandraDaemon.java:106) Exception encountered during startup: invalid column name length 0 On 10/10/11 9:24 AM, Sylvain Lebresne sylv...@datastax.com wrote: As said while closing previous vote, CASSANDRA-3338 felt important enough to warrant a reroll. So here it is: I propose the following artifacts for release as 1.0.0 (those are the same as the previous + the patch for
Re: [VOTE] Release Apache Cassandra 1.0.0 (Final) - Strike 2
On Tue, Oct 11, 2011 at 8:29 AM, Sylvain Lebresne sylv...@datastax.com wrote: So I have created new artifacts, based on the previous ones but with the infamous 'echo $CLASSPATH' removed from cassandra.in.sh. And what I'm proposing is to simply switch to those last artifacts for the release instead of the initial ones, but without restarting a vote just for that. +1. I agree there is absolutely zero chance of this breaking anything. -Brandon
Re: [VOTE] Release Apache Cassandra 1.0.0 (Final) - Strike 2
+1 On Tue, Oct 11, 2011 at 8:32 AM, Brandon Williams dri...@gmail.com wrote: On Tue, Oct 11, 2011 at 8:29 AM, Sylvain Lebresne sylv...@datastax.com wrote: So I have created new artifacts, based on the previous ones but with the infamous 'echo $CLASSPATH' removed from cassandra.in.sh. And what I'm proposing is to simply switch to those last artifacts for the release instead of the initial ones, but without restarting a vote just for that. +1. I agree there is absolutely zero chance of this breaking anything. -Brandon -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of DataStax, the source for professional Cassandra support http://www.datastax.com
Re: REST API?
There are no plans to move to Maven for 2 reasons: 1) a majority of committers favour Ant; 2) if it aint broke, don't fix it. On 11 October 2011 03:33, Brian O'Neill b...@alumni.brown.edu wrote: Will do. I've picked up where Gary left off. It is good starting point, with a good mapping between REST and get/set/mutations. (kudos to Gary) I'll update it to accomodate any changes and see if I can add some tests on top of it. I may look to add in JAX-RS (on either Jersey or Apache CXF). We use it for all of our REST services, and it may provide a good abstraction layer that we can build on. Give me a couple days. I have to get back into the ant mentality. I've been doing maven too long. BTW -- Does anyone know if there are plans to move to maven? (Not trying to start a religious war, just curious. ;) -brian On Mon, Oct 10, 2011 at 10:06 PM, Jeremy Hanna jeremy.hanna1...@gmail.comwrote: Brian, If you end up doing something with the rest api and making it available/open source, please post again either here or on the user list. I think others would be interested and may contribute to it. Cheers, Jeremy On Oct 10, 2011, at 8:42 PM, Brian O'Neill wrote: Thanks Gary. Perfect. Checking it out now. Performance isn't much of a concern for us through the REST interface. We are using the Hadoop/PIG integration to do the heavy lifting. This will be mostly for reads and small number of writes. I'll definitely give this a try. Thanks again. I'll let you know how it turns out. -brian On Mon, Oct 10, 2011 at 9:35 PM, Gary Dusbabek gdusba...@gmail.com wrote: It turns out that it is pretty easy (or it was a year ago) to replace the native Cassandra transport with your own. I wrote about it on my blog (http://www.onemanclapping.org/2010/09/restful-cassandra.html), using REST as an example. On Mon, Oct 10, 2011 at 20:12, Brian O'Neill b...@alumni.brown.edu wrote: My team desperately needs a REST API for Cassandra. I saw the following: http://code.google.com/p/restish/ from http://crlog.info/2011/01/29/restish-wrapper-for-hectorcassandra-data-manipulation/ But it appears to have little activity and documentation. That lead me to start work on a contrib/rest module, but before I get to far I wanted to ask if there was any effort underway for a REST Server/API. If not, I'll continue developing the REST server. Any preference for a REST stack? (JAX-RS on Apache-CXF? Raw Servlets? Netty? etc.) Until I hear back, I'll continue with the JAX-RS / Apache CXF implementation I have cooking. -brian -- Brian ONeill Lead Architect, Health Market Science ( http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/ -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/ -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/
Re: REST API?
I have one based on restlet: https://github.com/stinkymatt/Helena On Mon, Oct 10, 2011 at 9:12 PM, Brian O'Neill b...@alumni.brown.edu wrote: My team desperately needs a REST API for Cassandra. I saw the following: http://code.google.com/p/restish/ from http://crlog.info/2011/01/29/restish-wrapper-for-hectorcassandra-data-manipulation/ But it appears to have little activity and documentation. That lead me to start work on a contrib/rest module, but before I get to far I wanted to ask if there was any effort underway for a REST Server/API. If not, I'll continue developing the REST server. Any preference for a REST stack? (JAX-RS on Apache-CXF? Raw Servlets? Netty? etc.) Until I hear back, I'll continue with the JAX-RS / Apache CXF implementation I have cooking. -brian -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/
RE: REST API?
Please take me OFF this email...thanks All the Best Declan -Original Message- From: Matt Kennedy [mailto:stinkym...@gmail.com] Sent: Tuesday, October 11, 2011 10:09 AM To: dev@cassandra.apache.org Subject: Re: REST API? I have one based on restlet: https://github.com/stinkymatt/Helena On Mon, Oct 10, 2011 at 9:12 PM, Brian O'Neill b...@alumni.brown.edu wrote: My team desperately needs a REST API for Cassandra. I saw the following: http://code.google.com/p/restish/ from http://crlog.info/2011/01/29/restish-wrapper-for-hectorcassandra-data-manipu lation/ But it appears to have little activity and documentation. That lead me to start work on a contrib/rest module, but before I get to far I wanted to ask if there was any effort underway for a REST Server/API. If not, I'll continue developing the REST server. Any preference for a REST stack? (JAX-RS on Apache-CXF? Raw Servlets? Netty? etc.) Until I hear back, I'll continue with the JAX-RS / Apache CXF implementation I have cooking. -brian -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/
Re: [VOTE] Release Apache Cassandra 1.0.0 (Final) - Strike 2
On Mon, Oct 10, 2011 at 11:24 AM, Sylvain Lebresne sylv...@datastax.com wrote: As said while closing previous vote, CASSANDRA-3338 felt important enough to warrant a reroll. So here it is: I propose the following artifacts for release as 1.0.0 (those are the same as the previous + the patch for CASSANDRA-3338 and a one line fix to the build file so that it picks the right version of snappy-java from maven central). Unfortunately, I think https://issues.apache.org/jira/browse/CASSANDRA-3346 is a regression that will warrant another re-roll. :( Third time's the charm? -Brandon
Re: [VOTE CLOSE] Release Apache Cassandra 1.0.0 (Final) - Strike 2
Alright, that sucks. But with 3338, 3346 and 3345 (that is kind of a regression with respect to what the changelog says), I'm closing this vote and will start a new in a jiff. Sorry everyone. -- Sylvain On Tue, Oct 11, 2011 at 4:53 PM, Brandon Williams dri...@gmail.com wrote: On Mon, Oct 10, 2011 at 11:24 AM, Sylvain Lebresne sylv...@datastax.com wrote: As said while closing previous vote, CASSANDRA-3338 felt important enough to warrant a reroll. So here it is: I propose the following artifacts for release as 1.0.0 (those are the same as the previous + the patch for CASSANDRA-3338 and a one line fix to the build file so that it picks the right version of snappy-java from maven central). Unfortunately, I think https://issues.apache.org/jira/browse/CASSANDRA-3346 is a regression that will warrant another re-roll. :( Third time's the charm? -Brandon
ByteBuffers and StorageProxy
My recent bug was that I was sending a zero length ByteBuffer (because I forgot to flip) for a column name. The problem I have is that the insert was accepted by the server. Should an exception be thrown? The end result of allowing the insert is that the server will not restart if the data is still in the commit log (and maybe later too, not sure).
Re: ByteBuffers and StorageProxy
On Tue, Oct 11, 2011 at 12:25 PM, Todd Burruss bburr...@expedia.com wrote: My recent bug was that I was sending a zero length ByteBuffer (because I forgot to flip) for a column name. The problem I have is that the insert was accepted by the server. Should an exception be thrown? The end result of allowing the insert is that the server will not restart if the data is still in the commit log (and maybe later too, not sure). You may not have hit this at the thrift layer, because the AbstractTypes have validation methods. Since StorageProxy is the expert interface having redundant checks might cause code bloat. Open a jira and put in the patch. The worse that can happen is it gets -1 ed.
Re: REST API?
To give everyone an update... I was able to take what Gary had and update it to run on trunk. I like the native integration, as opposed to layering it on top of Hector. It's working out well. I layered in JAX-RS to replace the hand parsing of the url, and the handlers. I have reads and writes working through the StorageProxy, but I think I'm going to raise it up one layer to take advantage of ThriftValidations. (but still using direct method invocation instead of the thift client) I added unit tests for the read/write of columns. I'm going to add a few other operations (add/drop keyspace, add/drop CF). Then it should be in a state where I can share it. -brian On Mon, Oct 10, 2011 at 10:06 PM, Jeremy Hanna jeremy.hanna1...@gmail.comwrote: Brian, If you end up doing something with the rest api and making it available/open source, please post again either here or on the user list. I think others would be interested and may contribute to it. Cheers, Jeremy On Oct 10, 2011, at 8:42 PM, Brian O'Neill wrote: Thanks Gary. Perfect. Checking it out now. Performance isn't much of a concern for us through the REST interface. We are using the Hadoop/PIG integration to do the heavy lifting. This will be mostly for reads and small number of writes. I'll definitely give this a try. Thanks again. I'll let you know how it turns out. -brian On Mon, Oct 10, 2011 at 9:35 PM, Gary Dusbabek gdusba...@gmail.com wrote: It turns out that it is pretty easy (or it was a year ago) to replace the native Cassandra transport with your own. I wrote about it on my blog (http://www.onemanclapping.org/2010/09/restful-cassandra.html), using REST as an example. On Mon, Oct 10, 2011 at 20:12, Brian O'Neill b...@alumni.brown.edu wrote: My team desperately needs a REST API for Cassandra. I saw the following: http://code.google.com/p/restish/ from http://crlog.info/2011/01/29/restish-wrapper-for-hectorcassandra-data-manipulation/ But it appears to have little activity and documentation. That lead me to start work on a contrib/rest module, but before I get to far I wanted to ask if there was any effort underway for a REST Server/API. If not, I'll continue developing the REST server. Any preference for a REST stack? (JAX-RS on Apache-CXF? Raw Servlets? Netty? etc.) Until I hear back, I'll continue with the JAX-RS / Apache CXF implementation I have cooking. -brian -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/ -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/ -- Brian ONeill Lead Architect, Health Market Science (http://healthmarketscience.com) mobile:215.588.6024 blog: http://weblogs.java.net/blog/boneill42/ blog: http://brianoneill.blogspot.com/
