On Sun, Jul 31, 2011 at 2:25 AM, Jason Baker wrote:
> When I run nodetool repair on a node on my 3-node cluster, I see 3 messages
> like the following:
> INFO [manual-repair-6d9a617f-c496-4744-9002-a56909b83d5b] 2011-07-30
> 18:50:28,464 AntiEntropyService.java (line 636) No neighbors to repair w
Hi,
Let’s say that I have 2 datacenters, a key is changed on both of my
datacenters in the exact same time (even in 1-2 seconds diff).
Datacenter #1 remove a column and Datacenter #2 add 2 new columns.
Is there any problem with consistency or Cassandra will handle this
situation easily.
Thanks
> Let’s say that I have 2 datacenters, a key is changed on both of my
> datacenters in the exact same time (even in 1-2 seconds diff).
>
> Datacenter #1 remove a column and Datacenter #2 add 2 new columns.
> Is there any problem with consistency or Cassandra will handle this
> situation easily.
Co
You could try the following:
i:20110728 {
tx1="va1",
tx2="va1",
tx3="va1",
tx4="va1",
tx5="va1",
tx6="va1",
}
The value could either be a blob / json pojo, or a reference off to
another row storing the columns representing the value.
Taking it further
any help? thanks!
On Fri, Jul 29, 2011 at 12:05 PM, Yan Chunlu wrote:
> and by the way, my RF=3 and the other two nodes have much more capacity,
> why does they always routed the request to node3?
>
> coud I do a rebalance now? before node repair?
>
>
> On Fri, Jul 29, 2011 at 12:01 PM, Yan Chun
First run nodetool move and then you can run nodetool repair. Before you run
nodetool move you will need to determine tokens that each node will be
responsible for. Then use that token to perform move.
--
View this message in context:
http://cassandra-user-incubator-apache-org.3065146.n2.nabble.c
is that okay to do nodetool move before a completely repair?
using this equation?
def tokens(nodes):
- for x in xrange(nodes):
- print 2 ** 127 / nodes * x
On Mon, Aug 1, 2011 at 1:17 AM, mcasandra wrote:
> First run nodetool move and then you can run nodetool repair. Before you
> ru
I am running 3 nodes and RF=3, cassandra v0.7.4
seems when disablegossip and disablethrift could keep node in pretty low
load. sometimes when the node repair doing "rebuilding sstable", I would
disable gossip and thrift to lower the load. not sure if I could disable
them in the whole procedure. tha
I created an issue and attached a patch:
https://issues.apache.org/jira/browse/CASSANDRA-2979
I was not sure if it would be better to handle it in NodeProbe or
StorageService..
Bye,
Norman
2011/7/31 Sylvain Lebresne :
> On Sun, Jul 31, 2011 at 2:25 AM, Jason Baker wrote:
>> When I run nodetoo
In one of our test clusters we had a damaged commit log disks in one of the
nodes.
We have replication factor = 2 in this cluster, and write with consistency
level = ONE. So we expected writes will not be affected by such an issue. But
what actually happened is that the client that was writing
How about using Snowflake to generate the transaction ids:
https://github.com/twitter/snowflake
From: Kent Narling [mailto:kent.narl...@gmail.com]
Sent: Thursday, July 28, 2011 5:46 PM
To: user@cassandra.apache.org
Subject: Using Cassandra for transaction logging, good idea?
Hi!
I am considerin
A couple of timeouts should have kicked in.
First the rpc_timeout on the server side should have kicked in and given the
client a (thrift) TimedOutException. Secondly a client side socket timeout
should be set so the client will timeout the socket. Did either of these appear
in the client side
if you disable gossip the node will appear down to others. This would stop the
repair starting. After repair has started it *may* still cause problems when
new streams start (it probably does not). If the node is down other nodes will
stop sending writes to it.
disable thrift will stop clients
Thanks Aaron. We will try to pull the logs and post them in this forum.
But what I don't understand is why the client should pause at all. We are
writing with CL.ONE, and the replication factor is 2. As far as we understand -
the client communicates with a certain node (any node for that matter)
> aaron suggested it's better to run node repair on every node then re-balance
> it.
That's me been cautious with other peoples data.
It looks like node 3 is overwhelmed. Try getting the move sorted.
Cheers
-
Aaron Morton
Freelance Cassandra Developer
@aaronmorton
http://www.t
Yup, it sounds like things may not have failed as their should. Do you have a
better definition of stuck ? Was the client waiting for a single request to
completed or was the client not cycling to another node ?
If there is some server log details out it may help understand what happened.
Al
You may have better luck on the brisk user group
http://groups.google.com/group/brisk-users or IRC #datastax-brisk on freenode
I would guess you can do a rolling upgrade to the existing nodes. But brisk has
it's own snitch (BriskSimpleSnitch) so it may not be possible.
Cheers
---
Sounds interesting.
Reading a bit on snowflake it seems a bit uncertain if it fulfills the A & B
criterias?
ie:
> A, eventually return all known transactions
> B, Not return the same transaction more than once
Also, any reflections on the general idea to use Cassandra like this?
It wou
I think so to that Brisk may require all the nodes to be upgraded to Brisk.
Are there any detailed instructions about how to configure cassandra with
hadoop?
--
View this message in context:
http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Brisk-and-Hadoop-question-tp6638681p6639
springrider wrote:
>
> is that okay to do nodetool move before a completely repair?
>
> using this equation?
> def tokens(nodes):
>
>- for x in xrange(nodes):
> - print 2 ** 127 / nodes * x
>
Yes use that logic to get the tokens. I think it's safe to run move first
and reair later.
Check out http://wiki.apache.org/cassandra/HadoopSupport#ClusterConfig and that
whole page to see an intro to configuring your cluster. Brisk extends these
basic ideas.
On Jul 31, 2011, at 12:31 PM, mcasandra wrote:
> Is it possible to add brisk nodes for analytics to already existing real tim
On Saturday, July 30, 2011, Rafael Almeida wrote:
> Hello,
>
> I have computers that are better than others in my cluster. In special,
> there's one which is much better and I'd like to give it more load than the
> others. Is it possible? I'm using RandomPartitioner, should I use other?
> Should
I'm going to tell you guys the answers I could find so far.
On Tuesday, July 26, 2011, Rafael Almeida wrote:
> I couldn't find much documentation regarding how to make a cluster, but it
> seemed simple enough. At cassandra server A (10.0.0.2) I had seeds:
> "locahost". At server B (10.0.0.3) I
If you are doing insert only it should be ok. If you want a unique and roughly
ordered Tx id perhaps consider a TimeUUID in the first case, they are as
ordered as the clocks generating the UUID's. Which is about as good as
snowflake does, cannot remember what resolution the two use.
Be aware
The recommended approach is for all nodes in a cassandra cluster to have the
same HW spec. If the do not then you need to treat every node as having the
lowest possible spec (i.e. the lowest memory, lowest CPU, lowest disk capacity
and throughput). Other than during a HW upgrade, running mixed H
The wiki has info on setting up a cluster, see
http://wiki.apache.org/cassandra/Operations and
http://wiki.apache.org/cassandra/GettingStarted
If get errors check the server side logs (/var/log/cassandra), also make sure
that you are getting the exception raised by thrift. e.g.
TInvalidRequest
On Mon, Aug 1, 2011 at 8:24 AM, Rafael Almeida wrote:
> On Saturday, July 30, 2011, Rafael Almeida wrote:
> > Hello,
> >
> > I have computers that are better than others in my cluster. In special,
> > there's one which is much better and I'd like to give it more load than
> the
> > others. Is i
Hi,
I was wondering if anyone would know if secondary index can be enabled on
composite columns?
Regards
Boris
okay, I see. thanks a lot for the help!
On Mon, Aug 1, 2011 at 5:26 AM, aaron morton wrote:
> if you disable gossip the node will appear down to others. This would stop
> the repair starting. After repair has started it *may* still cause problems
> when new streams start (it probably does not). I
okay, thanks Aaron!
On Mon, Aug 1, 2011 at 5:43 AM, aaron morton wrote:
> aaron suggested it's better to run node repair on every node then
> re-balance it.
>
>
> That's me been cautious with other peoples data.
>
> It looks like node 3 is overwhelmed. Try getting the move sorted.
>
> Cheers
>
>
thanks a lot! I will try the "move".
On Mon, Aug 1, 2011 at 7:07 AM, mcasandra wrote:
>
> springrider wrote:
> >
> > is that okay to do nodetool move before a completely repair?
> >
> > using this equation?
> > def tokens(nodes):
> >
> >- for x in xrange(nodes):
> > - print 2 ** 127 /
Hi, my read latency is really horrible and I can't figure out what went
wrong. I'm running cassandra 0.8.0 on a 5 machine cluster. The Fingerprint
ColumnFamily has 400,000 rows, each row has about 4,000 Super columns, and
each super column has 1 to 4 columns. One row looks like:
RowKey: 00c26f
Sure, but it's still only useful for equality predicates.
On Sun, Jul 31, 2011 at 8:50 PM, Boris Yen wrote:
> Hi,
> I was wondering if anyone would know if secondary index can be enabled on
> composite columns?
> Regards
> Boris
--
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of
Hi,
try running a major compaction via nodetool on this Column family. The number
of SSTables seems quite large.
Considering the space used, this might take a few hours and might also impact
performance.
Cheers,
T.
On 01/08/11 14:23, myreasoner wrote:
Hi, my read latency is reall
If I do
./nodetool -h localhost compact keyspace columnfamily1
it will go out and compact coumnfamily1 on all the nodes not just the
localhost, correct?
--
View this message in context:
http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Read-latency-is-over-1-minute-on-a-column-fa
Compaction is machine-local, you need to run it on every node. Do it as a
rolling compaction (or in parallel if you can take the performance hit).
Cheers,
T.
On 01/08/11 15:31, myreasoner wrote:
If I do
./nodetool -h localhost compact keyspace columnfamily1
it will go out and compac
Thanks.
I did *./nodetool -h localhost compact keyspace columnfamily1 *. But it
came back really quick and the cfstats doesn't seem change much.
After compaction:
Column Family: Fingerprint
SSTable count: 2057
Space used (live): 164351343468
Hi,
try
nodetool -h localhost compact
check progress with
nodetool -h localhost compactionstats
and check system.log
Cheers,
T.
On 01/08/11 15:47, myreasoner wrote:
Thanks.
I did *./nodetool -h localhost compact keyspace columnfamily1 *. But it
came back really quick and the cfs
Did you run that verbatim ? Or you appropriately substituted "keyspace" and
"columnfamily1" ?
Also, anything in cassandra's log file (system.log) ? Compacting 150Gb over
2057 SSTables should take a reasonable bit of time...
On 2011-07-31, at 11:47 PM, myreasoner wrote:
> Thanks.
>
> I did
On the node that the compaction returned almost immediately:
*woot@n50:~$ /opt/cassandra/bin/nodetool -h localhost compactionstats
pending tasks: 66*
However, messages shown on other nodes are:
compaction type: Major
keyspace: MyKeyspace
column family: Fingerprint
bytes compacted: 25505066421
byt
Looks like a broken node, just restart Cassandra on that node. Might want to
wait for the compaction to finish on the other nodes.
Also, don't forget to JMX gc() manually after the compaction has finished to
delete the files on each node.
On 01/08/11 16:29, myreasoner wrote:
On the node that
41 matches
Mail list logo
