On Fri, Dec 12, 2014 at 4:50 PM, Eric Stevens wrote:
>
> I know that Thrift includes keyspace as part of the connection details, so
> if you're reading or writing to many keyspaces, you'll end up having to
> make a lot of additional round trips, and it will hurt your throughput. I
> may be wrong
Well we started with the thought that we'd have two keyspaces, one for
searchables and one for non-searchables like you mentioned. But our
concern is that we may change our mind about what column families are
available for search in the future. Separate keyspaces per table give us
greater flexibi
You can also collect these stats from the server via JMX, I believe the name of
the MBean object is
org.apache.cassandra.metrics:type=ColumnFamily,keyspace=,scope=,name=SSTablesPerReadHistogram
where KEYSPACE is your keyspace and SCOPE is Read or Write.
This has attributes for 50thPercentile, 7
well did you restart cassandra after changing the JVM_OPTS to match your
desired address?
On Fri, Dec 12, 2014 at 2:34 PM, Kevin Burton wrote:
>
> Oh. and if I specify —host it still doesn’t work. Very weird.
>
> On Fri, Dec 12, 2014 at 12:33 PM, Kevin Burton wrote:
>
>> OK..I’m stracing it and
OK..I’m stracing it and it’s definitely trying to connect to 173… here’s
the log line below. (anonymized).
the question is why.. is cassandra configured to return something on the
public address via JMX? I guess I could dump all of JMX metrics and figure
it out.
[pid 32331] connect(41, {sa_famil
Oh. and if I specify —host it still doesn’t work. Very weird.
On Fri, Dec 12, 2014 at 12:33 PM, Kevin Burton wrote:
> OK..I’m stracing it and it’s definitely trying to connect to 173… here’s
> the log line below. (anonymized).
>
> the question is why.. is cassandra configured to return somethi
hmm I was hoping it was changed in 2.1
https://github.com/apache/cassandra/blob/cassandra-2.1/src/java/org/apache/cassandra/tools/NodeTool.java
but still localhost, sorry I can't tell you why it would go to the public
interface..maybe someone added a shell alias?
On Fri, Dec 12, 2014 at 2:20 PM,
is appears to be localhost, I imagine the issue is more you changed the
rpc_address to not be localhost anymore
https://github.com/apache/cassandra/blob/cassandra-2.0/src/java/org/apache/cassandra/tools/NodeCmd.java
lines 87 and 88
private static final String DEFAULT_HOST = "127.0.0.1";
private s
AH! … ok. I didn’t see that nodetool took a host. Hm.. How does it
determine the host to read from by default?
The problem is that somehow it wants to read from the public interface
(which is fire walled)
On Fri, Dec 12, 2014 at 5:19 AM, Ryan Svihla wrote:
> yes the node needs to restart to ha
On Tue, Dec 9, 2014 at 11:26 PM, Jonathan Haddad wrote:
> Yes. It is, in general, a best practice to upgrade to the latest bug fix
> release before doing an upgrade to the next point release.
AFAIK, no one is formally testing upgrades from the middle of a series,
so...
+1, users should consid
Hello there -
I'm using the python-driver to get my queried rows with a row factory
that's a dictionary. When I get back my row list, each list item is a
dictionary, but the keys are hashes. Is there any way use the column family
metadata to decrypt the dictionary keys to their original column nam
Hey Jens,
Unfortunately the output of the nodetool histograms changes between
versions. While I think your script is useful, it's likely to break
between versions. You might be interested to weigh in on the JIRA ticket
to make the nodetool output machine friendly:
https://issues.apache.org/jira/
The really important thing to really take away from Ryan's original post is
that batches are not there for performance. The only case I consider
batches to be useful for is when you absolutely need to know that several
tables all get a mutation (via logged batches). The use case for this is
when
It would make more sense to just have a keyspace for each. Something like
solr_tables, and cassandra_tables. I've done similar with most customers
using DSE search (not a DSE mailing list, but the information is
interesting background for your question).
there is a cost to each keyspace and you'll
Clarification "keyspace for each" should be "keyspace for cassandra tables
and solr tables"
On Fri, Dec 12, 2014 at 11:25 AM, Ryan Svihla wrote:
>
> It would make more sense to just have a keyspace for each. Something like
> solr_tables, and cassandra_tables. I've done similar with most customers
We're considering moving to a model where we put each of our tables in a
dedicated keyspace. This is so we can tune replication per table, and
change our mind about that replication on a per-table basis without a major
migration. The biggest driver for this is Solr integration, we want to
tune RF
Hi,
I just quickly put together a tiny utility script to estimate
average/mean/min/max/percentiles for `nodetool cfhistogram` latency output.
Maybe could be useful to someone else, don’t know. You can find it here:
https://gist.github.com/JensRantil/3da67e39f50aaf4f5bce
Future improvements
yes the node needs to restart to have cassandra-env.sh take effect, and the
links you're providing are about making cassandra's JMX bind to the
interface you want, so nodetool isn't really the issue, nodetool can just
take an ip argument to connect to the interface you desire.Something like:
nodet
Any insert, update, or delete
On Fri, Dec 12, 2014 at 1:31 AM, Jens Rantil wrote:
>
> Maybe slightly off-topic, but what is a mutation? Is it equivalent to a
> CQL row? Or maybe a column in a row? Does include tombstones within the
> selected range?
>
> Thanks,
> Jens
>
>
>
> On Thu, Dec 11, 2014
It's a rough observation and estimate, nothing more. In other words, some
clusters can handle more, some can't, it depends on how many writes per
second you're doing, cluster sizing, how far over that 5kb limit you are,
heap size, disk IO, cpu speed, and many more factors. This is why it's just
a w
What version are you on (key estimate I see in 1.2 and 2.0) ? What size is
your heap (ideally 8GB, can be lower, but it requires a lot of tuning)?
What kind of disk do you have (SANs are going to cause you problems)?
Assuming all of those are the right answer, then you have the following
options to
21 matches
Mail list logo