Before you dig into that, are you sure normal Thrift inserts are not
fast enough?
On Thu, Nov 11, 2010 at 4:41 PM, Aditya Muralidharan
wrote:
> Pretty sure I could ask that better:
>
>
>
> Is it possible for me to perform RowMutations on BinaryMemtable for a
> ColumnFamily of type Super?
>
>
>
>
I've been running tests with a first four-node, then eight-node
cluster. I started with 0.7.0 beta3, but have since updated to a more
recent Hudson build. I've been happy with a lot of things, but I've had
some really surprisingly unpleasant experiences with operational fragility.
For example, w
Thanks, Peter.
I missed an important point. Each time the performance resets, there
is a memtable flush. The memtable is flushed every hour. It is about
128MB. Each time the memtable flushed, there is a GC spike.
After a memtable flush, you see minimum cpu and maximum read
throughput both in term
> - About 300 million records are stored in single cassandra node. (16
> cores, 32GB memory, SSD disk)
> - One record is avg. 1KB.
So about 300 GB of disk space in total?
> - Multi-thread client on different server does read and write:
> - 120 reader threads continuously read random records from
Pretty sure I could ask that better:
Is it possible for me to perform RowMutations on BinaryMemtable for a
ColumnFamily of type Super?
The bmt_example seems to say that it's possible, but cassandra 0.7 b3 seems to
disagree with the following:
ERROR [MutationStage:38] 2010-11-11 13:47:37,383
D
Can you try with the latest 0.6 svn branch?
https://svn.apache.org/repos/asf/cassandra/branches/cassandra-0.6 It
may be caused by https://issues.apache.org/jira/browse/CASSANDRA-1722
which is fixed there.
On Thu, Nov 11, 2010 at 4:56 AM, Eric van Orsouw
wrote:
> I have tried it with RF=1 and 2 n
On 11/11/10 1:34 PM, Robert wrote:
Brandon Williams gmail.com> writes:
As the wiki says: "The latency of operations since the last time this
attribute was read."
You mean, the last time any client queried the JMX MBean attribute?
Yes.
=Rob
Brandon Williams gmail.com> writes:
> As the wiki says: "The latency of operations since the last time this
attribute was read."
You mean, the last time any client queried the JMX MBean attribute?
Is it possible for BinaryMemtable RowMutations to a ColumnFamily with
supercolumns?
On Thu, Nov 11, 2010 at 2:53 PM, Robert <
cassandra.robertw...@spamgourmet.com> wrote:
> I'm looking at
>
> * RecentRangeLatencyMicros
> * RecentReadLatencyMicros
> * RecentWriteLatencyMicros
>
> on http://wiki.apache.org/cassandra/JmxInterface . However, it is not
> clear
> (short of looking
I'm looking at
* RecentRangeLatencyMicros
* RecentReadLatencyMicros
* RecentWriteLatencyMicros
on http://wiki.apache.org/cassandra/JmxInterface . However, it is not clear
(short of looking at the source code) what "recent" means.
Does anybody know? I've tried running some stuff and querying
It's working as written, but I think you're right that it makes more
sense to fail the expression when the column doesn't exist.
On Thu, Nov 11, 2010 at 7:04 AM, Ching-Cheng Chen
wrote:
> Not sure if this the intended behavior of the indexed query.
>
> I created a column family and added index on
My understanding is the clauses are AND'd together. There is some nice logging from the ColumnFamilyStore when it's running the request, can you turn the logging up to DEBUG and send them along? The first log message will say something like "Primary scan clause is.." CheersAaronOn 12 Nov, 2010,at 0
Do you have some existing performance issues that you are trying to resolve? It's easier to improve performance if you know you have X nodes and Y requests.Each write will be sent to all replicas that are up, then number is determined from your RF (set RF to all nodes in the cluster to store the da
On Thu, 2010-11-11 at 14:00 -0500, Sam Hendley wrote:
> Thanks for the update and the especially the debian packages. I have 2
> debian package related questions:
>
> 1) The keyserver wwwkeys.eu.pgp.net seems to go down for a few days at
> a time. That can make installing to a new box impossible u
On Thu, Nov 11, 2010 at 1:00 PM, Sam Hendley
wrote:
> Thanks for the update and the especially the debian packages. I have 2
> debian package related questions:
>
> 1) The keyserver wwwkeys.eu.pgp.net seems to go down for a few days at a
> time. That can make installing to a new box impossible unt
Thanks for the update and the especially the debian packages. I have 2
debian package related questions:
1) The keyserver wwwkeys.eu.pgp.net seems to go down for a few days at a
time. That can make installing to a new box impossible until it comes up. I
have created a gist with the key in the mean
On Thu, Nov 11, 2010 at 4:56 AM, Eric van Orsouw wrote:
> I appears to me there is an inconsistency somewhere, I hope this info
> helps.
get_range_slices does not perform read repair, so inconsistencies at cl.ONE
won't be resolved. Invoke nodetool repair everywhere and try again.
-Brandon
Done, thanks for the reminder.
On Thu, Nov 11, 2010 at 12:15 PM, gbanks wrote:
> sorry for the empty body... i hit send a bit prematurely.
>
> so, rename functionality is gone until
> https://issues.apache.org/jira/browse/CASSANDRA-1585 is resolved?
>
> in that case, can i also suggest removing:
sorry for the empty body... i hit send a bit prematurely.
so, rename functionality is gone until
https://issues.apache.org/jira/browse/CASSANDRA-1585 is resolved?
in that case, can i also suggest removing:
rename keyspace Rename a keyspace.
rename column family
At first glance, this appeared to be a very egregious bug, but the effect is
actually minimal: since the size of the buffer is deterministic based on the
size of the data, you will have equal amounts of excess/junk data for equal
rows. Combined with the fact that 0.6 doesn't reuse these buffers,
Not sure if this the intended behavior of the indexed query.
I created a column family and added index on column A,B,C.
Now I insert three rows.
row1 : A=123, B=456, C=789
row2 : A=123, C=789
row3 : A=123, B=789, C=789
Now if I perform an indexed query for A=123 and B=456, both row1 and row2
ar
Hi,
thanks for all the informative answers.
Since writing is much faster then reading, I assume that's it's faster to
write the data to 3 replicas and read from 1 instead of writing to 2 and
reading from at least 2. (especially if I execute the read operation
multiple times on the same key). I co
I have tried it with RF=1 and 2 nodes and the result was the same.
In this case I am not able to query anything, regardless the consistency level.
I have conducted some more tests with 4 nodes and RF=2 and before each test I
also completely cleared all datafiles (data/commit/saved_cache).
Each
Hi JE,
0.6.6:
org.apache.cassandra.service.AntiEntropyService
I found the rowHash method uses "row.buffer.getData()" directly.
Since row.buffer.getData() is a byte[], and there may have some junk bytes
in the end by the buffer, I think we should use the exact length.
private MerkleTree.
25 matches
Mail list logo