Hello list,
I have a problem with my cluster ownership not being as expected.
I have 2 DC cluster using NetworkTopologyStrategy on and
EC2MultiRegionSnitch with cassandra 1.1.5. My placement strategy for all
keyspaces is: {eu-west: 3, us-east:3 }, and I have 6 nodes in eu-west and 3
in us-east.
Hello Katriel,
Are you using replication factor 3? So it seems that in us-east it's all
fine with each machine having 100% of the data. The problem is in eu-west
where you have unbalanced nodes. You should rearrange the tokens in that
eu-west DC.
Check e.g. this token calculator to calculate
The problem is that NetworkTopologyStrategy will try to pick nodes that
have a different rack when going around the ring, so the second node in
each rack always gets skipped unless it was the first node picked. Your
nodes is eu-west go a,a,b,b,c,c but they should be a,b,c,a,b,c.
On Wed, Nov 13,
Thats it! I knew NetworkTopologyStrategy goes through racks when writing,
but didn't consider that the first one will most likely be used.
I'll re-arrange my nodes to be a,b,c,a,b,c
Thanks,
Katriel
On Wed, Nov 13, 2013 at 3:15 PM, Derek Williams de...@fyrie.net wrote:
The problem is that
You basically want option (c). Option (d) might work, but you would be
bending the paradigm a bit, IMO. Certainly do not use dedicated column
families or keyspaces per tennant. That never works. The list history will
show that with a few google searches and we've seen it fail badly with
several
Nate,
(slightly OT), what client API/library is recommended now that Hector is
sunsetting? Thanks.
Regards,
Shahab
On Wed, Nov 13, 2013 at 9:28 AM, Nate McCall n...@thelastpickle.com wrote:
You basically want option (c). Option (d) might work, but you would be
bending the paradigm a bit,
Astyanax and/or the DS Java client depending on your use case. (Emphasis on
the and - really no reason you can't use both - even on the same schema -
depending on what you are doing as they both have their strengths and
weaknesses).
To be clear, Hector is not going away. We are still accepting
On Wed, Nov 13, 2013 at 5:34 AM, Katriel Traum katr...@google.com wrote:
Thats it! I knew NetworkTopologyStrategy goes through racks when writing,
but didn't consider that the first one will most likely be used.
I'll re-arrange my nodes to be a,b,c,a,b,c
On Wed, Nov 13, 2013 at 12:35 AM, Tom van den Berge t...@drillster.comwrote:
I'm having the same problem, after upgrading from 1.2.3 to 1.2.10.
I can remember this was a bug that was solved in the 1.0 or 1.1 version
some time ago, but apparently it got back.
A workaround is to delete the
Hello,
I was able to run WordCount in $CASSANDRA_HOME/examples/hadoop_word_count and
the Cassandra version is 2.0.2
So when I start the word_count, the actual process is a Java -cp . Is there
a Jobtracker started by the Cassandra or how is the wordcount mapreduce job
run. Is there a task
IIRC there is a test for snappy when the node starts does that log an error ?
And / or can you create a CF that uses snappy compression (it was the default
for a while in 1.2).
Cheers
-
Aaron Morton
New Zealand
@aaronmorton
Co-Founder Principal Consultant
Apache Cassandra
11 matches
Mail list logo
