RE: Why is Quorum not sufficient for Linearization?

2014-10-16 Thread Ruebenacker, Oliver A
Hello, The fact that things can always change immediately is not an obstacle to linearizability. The lack of linearizability manifests in inconsistency, i.e. you read from multiple nodes and get different results. What does Cassandra do in the case of inconsistent reads? Wait or

Re: Why is Quorum not sufficient for Linearization?

2014-10-16 Thread Peter Lin
To the best of my knowledge, only guaranteed way is with an ACID compliant system. The examples other have already provided should give you a decent idea. If that's not enough, you would need to read papers on CRDT's and how they compare to ACID systems.

RE: validation compaction

2014-10-16 Thread S C
Bob, Default compression is Snappy compression and I have seen compression ranging between 2-4% (just as the doc says). I got the storage part. Does it mean that as a result of compaction/repair SSTables are decompressed? Is it the reason for CPU utilization spiking up a little? -SR From:

Re: Why is Quorum not sufficient for Linearization?

2014-10-16 Thread Evan Weaver
Quorum reads and writes in Cassandra guarantee sequential consistency. The reason this doesn't satisfy linearizability is because resurrections of unacknowledged writes can occur. A read of a half-committed write will trigger synchronous read repair and the order will be stable from that point

are all files for an sstable immutable?

2014-10-16 Thread Sean Bridges
Hello, I thought an sstable was immutable once written to disk. Before upgrading from 1.2.18 to 2.0.10 we took a snapshot of our sstables. Now when I compare the files in the snaphot dir and the original files, the Summary.db files have a newer modified date, and the file sizes have changed.

Re: are all files for an sstable immutable?

2014-10-16 Thread Tyler Hobbs
The summary files are immutable, but can be replaced periodically. See https://issues.apache.org/jira/browse/CASSANDRA-5519 for more details. The summary files aren't particularly important, they're primarily an optimization for startup time. On Thu, Oct 16, 2014 at 12:20 PM, Sean Bridges

Re: are all files for an sstable immutable?

2014-10-16 Thread Sean Bridges
Thanks Tyler, that explains it. Sean On Thu, Oct 16, 2014 at 10:40 AM, Tyler Hobbs ty...@datastax.com wrote: The summary files are immutable, but can be replaced periodically. See https://issues.apache.org/jira/browse/CASSANDRA-5519 for more details. The summary files aren't particularly

Re: are all files for an sstable immutable?

2014-10-16 Thread Robert Coli
On Thu, Oct 16, 2014 at 10:40 AM, Tyler Hobbs ty...@datastax.com wrote: The summary files are immutable, but can be replaced periodically. See https://issues.apache.org/jira/browse/CASSANDRA-5519 for more details. The summary files aren't particularly important, they're primarily an

Re: validation compaction

2014-10-16 Thread Robert Coli
On Thu, Oct 16, 2014 at 6:41 AM, S C as...@outlook.com wrote: Bob, Bob is my father's name. Unless you need a gastrointestinal consult, you probably don't want to ask Bob Coli a question... ;P Default compression is Snappy compression and I have seen compression ranging between 2-4% (just

Re: What will be system configuration for retrieving few GB of data

2014-10-16 Thread Robert Coli
On Wed, Oct 15, 2014 at 10:46 PM, Umang Shah shahuma...@gmail.com wrote: I am facing many problem after storing certain limit of records in cassandra, and giving outofmemoryerror. This description is too vague. Be more specific. =Rob http://twitter.com/rcolidba

Re: Should a node that is bootstrapping be receiving writes in addition to the streams it is receiving?

2014-10-16 Thread Robert Coli
On Wed, Oct 15, 2014 at 10:07 PM, Peter Haggerty peter.hagge...@librato.com wrote: The node wrote gigs of data to various CFs during the bootstrap so it was clearly writing in some sense and it has the expected behavior after the bootstrap. Is cfstats correct when it reports that there were

Re: Question about adding nodes incrementally to a new datacenter: wait til all hosts come up so they can learn the token ranges?

2014-10-16 Thread Robert Coli
On Wed, Oct 15, 2014 at 3:25 PM, Donald Smith donald.sm...@audiencescience.com wrote: So, my point is that to avoid the need to bootstrap and to cleanup, it's better to bring all nodes up at about the same time. If this is wrong, please explain why. Oh, sure. As you say, you avoid having

Deleting data in 1.2.8 based on time stamp

2014-10-16 Thread Rahul Neelakantan
Does anyone know of a way to delete rows from C* 1.2.8 based on the timestamp (time from epoch) that is present on each column in the triplet of name, value and timestamp? (I do not have a separate date/timestamp column that I insert) Rahul Neelakantan

Re: Deleting data in 1.2.8 based on time stamp

2014-10-16 Thread Tyler Hobbs
For each partition in the table, run: DELETE FROM mytable WHERE partitionkey=? USING TIMESTAMP 123456789 And it will delete everything older than or equal to 123456789 (in microseconds since the epoch, if you're using standard timestamps). On Thu, Oct 16, 2014 at 5:09 PM, Rahul Neelakantan

[Release] Achilles 3.0.7

2014-10-16 Thread DuyHai Doan
Hello Cassandra users We are happy to announce the release of Achilles 3.0.7, an advanced object mapper built upon the Java driver. Apart from usual bug fixes and perf improvement: - support for multi-keyspaces - support for naming strategies (snake case, case sensitive lower case for

Re: Deleting data in 1.2.8 based on time stamp

2014-10-16 Thread Rahul Neelakantan
So this would need me to know the partition keys, what if I simply wanted to say delete all rows where the timestamp was older than 123456789? Rahul Neelakantan On Oct 16, 2014, at 6:27 PM, Tyler Hobbs ty...@datastax.com wrote: For each partition in the table, run: DELETE FROM mytable

Re: Deleting data in 1.2.8 based on time stamp

2014-10-16 Thread Tyler Hobbs
On Thu, Oct 16, 2014 at 5:32 PM, Rahul Neelakantan ra...@rahul.be wrote: So this would need me to know the partition keys, what if I simply wanted to say delete all rows where the timestamp was older than 123456789? You can't. You'll need to loop over the table and collect the keys. --

Re: CQL Alter table does not propagate to all nodes.

2014-10-16 Thread dlu66061
Thank you very much, Erick. Yes, we are using NTP. But your other suggestions and links are very helpful. I tried to grep MigrationStage from system.log and found Can't send migration request: node /201.20.32.54 is down. around the time I ran the CQL, although that server is actually up running

Re: How to restore a snapshot in cassandra 1.1

2014-10-16 Thread Ben Chobot
We're wiping the commit logs because that's what the datastax instructions say to do. (Also the old Cassandra Ops wiki.) I assume it's so that changes that no longer apply to the node aren't replayed when it's restarted with old sstables. Of course, my question is about when you have multiple

Cassandra Restore data from snapshots and Different Counts

2014-10-16 Thread Bosung Seo
I upgraded my Cassandra ring and restored data(copying snapshots) from the old ring. I am currently running the nodetool repair. I count the tables to check every rows is in the table, but counts have different values. It contains 571 rows, and counts are 500, 530, 501, and so on. Should I wait

Re: Cassandra Restore data from snapshots and Different Counts

2014-10-16 Thread Robert Coli
On Thu, Oct 16, 2014 at 4:17 PM, Bosung Seo bos...@brightcloud.com wrote: I upgraded my Cassandra ring and restored data(copying snapshots) from the old ring. I am currently running the nodetool repair. I count the tables to check every rows is in the table, but counts have different values.

Re: How to restore a snapshot in cassandra 1.1

2014-10-16 Thread Robert Coli
On Thu, Oct 16, 2014 at 3:57 PM, Ben Chobot be...@instructure.com wrote: We're wiping the commit logs because that's what the datastax instructions say to do. (Also the old Cassandra Ops wiki.) I assume it's so that changes that no longer apply to the node aren't replayed when it's restarted