incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From shiv shivaji <shivaji...@yahoo.com>
Subject Re: Anti-compaction Diskspace issue even when latest patch applied
Date Wed, 03 Mar 2010 01:26:13 GMT
Thanks, just realized this after looking at the source code.

Seems like the decommission will not work for me due to disk space issues. I am currently
moving all the data on the heavy node (5 TB full) to a 12 TB disk drive. I am planning to
remove the old token and resign the old token to this node.

According to the docs, it says to use decommission, however lack of disk space does not allow
me to do this. If I manually move all the data files and then do a removetoken and start the
node with a new token, would that work (as was implied in a JIRA)?

Shiv





________________________________
From: Stu Hood <stu.hood@rackspace.com>
To: cassandra-user@incubator.apache.org
Sent: Sun, February 28, 2010 1:53:29 PM
Subject: Re: Anti-compaction Diskspace issue even when latest patch applied

`nodetool cleanup` is a very expensive process: it performs a major compaction, and should
not be done that frequently.

-----Original Message-----
From: "shiv shivaji" <shivajisus@yahoo.com>
Sent: Sunday, February 28, 2010 3:34pm
To: cassandra-user@incubator.apache.org
Subject: Re: Anti-compaction Diskspace issue even when latest patch applied

Seems like the temporary solution was to run a cron job which calls nodetool cleanup every
5 mins or so. This stopped the disk space from going too low.

The manual solution you mentioned is likely worthy of consideration as the load balancing
is taking a while.

I will track the jira issue of anticompaction and diskspace. Thanks for the pointer.


Thanks, Shiv




________________________________
From: Jonathan Ellis <jbellis@gmail.com>
To: cassandra-user@incubator.apache.org
Sent: Wed, February 24, 2010 11:34:59 AM
Subject: Re: Anti-compaction Diskspace issue even when latest patch applied

as you noticed, "nodeprobe move" first unloads the data, then moves to
the new position.  so that won't help you here.

If you are using replicationfactor=1, scp the data to the previous
node on the ring, then reduce the original node's token so it isn't
responsible for so much, and run cleanup.  (you can do this w/ higher
RF too, you just have to scp the data more places.)

Finally, you could work on
https://issues.apache.org/jira/browse/CASSANDRA-579 so it doesn't need
to anticompact to disk before moving data.

-Jonathan

On Wed, Feb 24, 2010 at 12:06 PM, shiv shivaji <shivajisus@yahoo.com> wrote:
> According to the stack trace I get in the log, it makes it look like the
> patch was for anti-compaction but I did not look at the source code in
> detail yet.
>
> java.util.concurrent.ExecutionException:
> java.lang.UnsupportedOperationException: disk full
>         at
> java.util.concurrent.FutureTask$Sync.innerGet(FutureTask.java:222)
>         at java.util.concurrent.FutureTask.get(FutureTask.java:83)
>         at
> org.apache.cassandra.concurrent.DebuggableThreadPoolExecutor.afterExecute(DebuggableThreadPoolExecutor.java:86)
>         at
> org.apache.cassandra.db.CompactionManager$CompactionExecutor.afterExecute(CompactionManager.java:570)
>         at
> java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:888)
>         at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
>         at java.lang.Thread.run(Thread.java:619)
> Caused by: java.lang.UnsupportedOperationException: disk full
>         at
> org.apache.cassandra.db.CompactionManager.doAntiCompaction(CompactionManager.java:344)
>         at
> org.apache.cassandra.db.CompactionManager.doCleanupCompaction(CompactionManager.java:405)
>         at
> org.apache.cassandra.db.CompactionManager.access$400(CompactionManager.java:49)
>         at
> org.apache.cassandra.db.CompactionManager$2.call(CompactionManager.java:130)
>         at
> java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)
>         at java.util.concurrent.FutureTask.run(FutureTask.java:138)
>         at
> java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
>         ... 2 more
>
> I tried "nodetool cleanup" before and it did not really stop the disk from
> filling, is there a way to force move the data or some other way to solve
> the issue?
>
> Thanks, Shiv
>
> ________________________________
> From: Jonathan Ellis <jbellis@gmail.com>
> To: cassandra-user@incubator.apache.org
> Sent: Wed, February 24, 2010 7:16:32 AM
> Subject: Re: Anti-compaction Diskspace issue even when latest patch applied
>
> The patch you refer to was to help *compaction*, not *anticompaction*.
>
> If the space is mostly hints for other machines (is that what you
> meant by "due to past problems with others?") you should run nodeprobe
> cleanup on it to remove data that doesn't actually belong on that
> node.
>
> -Jonathan
>
> On Wed, Feb 24, 2010 at 3:09 AM, shiv shivaji <shivajisus@yahoo.com> wrote:
>> For about 6TB of  total data size with a replication factor of 2 (6TB x 2)
>> on a five node cluster, I see about 4.6 TB on one machine (due to
>> potential
>> past problems with other machines). The machine has a disk of 6TB.
>>
>> The data folder on this machine has 59,289 files totally 4.6 TB. The files
>> are the data, filter and indexes. I see that anti-compaction is running. I
>> applied a recent patch which does not do anti-compaction if disk space is
>> limited. I still see it happening. I have also called nodetool loadbalance
>> on this machine. Seems like it will run out of disk space anyway.
>>
>> The machine diskspace consumed are: (Each machine has a 6TB hard-drive on
>> RAID).
>>
>> Machine Space Consumed
>> M1    4.47 TB
>> M2    2.93 TB
>> M3    1.83 GB
>> M4    56.19 GB
>> M5    398.01 GB
>>
>> How can I force M1 to immediately move its load to M3 and M4 for instance
>> (or any other machine). The nodetool move command moves all data, is there
>> a
>> way instead to force move 50% of data to M3 and the remaining 50% to M4
>> and
>> resume anti-compaction after the move?
>>
>> Thanks, Shiv
>>
>>
>>
>
Mime
View raw message