See https://builds.apache.org/job/HBase-0.92/294/changes
Changes:
[stack] HBASE-5209 HConnection/HMasterInterface should allow for way to get
hostname of currently active master in multi-master HBase setup -- FIX BUILD ON
0.92; ADDENDUM
--
[...truncated
Hi,
Shall we experiment with low inter-reference recency set replacement policy to
see if block cache becomes more effective ?
Cheers
Hi Devs
We ran into one issue while splitting HLogs due to EOFException. (0.90.6
version).
Due to some reason the DNs were not able to connect to NN (network
fluctuation) and the master was splitting the logs.
While parsing the hlog, we get the length and we expect it might be 0. (No
Can you provide stack trace for this issue ?
Thanks
On Feb 21, 2012, at 8:52 AM, Ramkrishna.S.Vasudevan
ramkrishna.vasude...@huawei.com wrote:
Hi Devs
We ran into one issue while splitting HLogs due to EOFException. (0.90.6
version).
Due to some reason the DNs were not able
Please find the logs in HRegionServer. Did you mean this Ted?
2012-02-18 00:44:38,808 INFO org.apache.hadoop.hbase.util.FSUtils: Finished
lease recover attempt for
hdfs://158-1-130-13:9000/hbase/.logs/linux1,20020,1329487401169/linux1%3A200
20.1329492399793
2012-02-18 00:44:38,808 WARN
We had the author of LIRS come to Facebook last year to talk about his
algorithm and general benefits. At the time, we were looking at
increasing block cache efficiency. The general consensus was that it
wasn't an exponential perf gain, so we could get bigger wins from
cache-on-write
afaik, existing LruBlockCache is not exactly LRU cache
It utilizes more advanced algorithm to avoid cache trashing during scan ops by
dividing cache into three sub-caches (for newly added blocks, for promoted
blocks and for in-memory blocks)
Best regards,
Vladimir Rodionov
Principal Platform
I thought about this over the summer when I was working on 4027.
Pretty much same idea as Nicholas here. I figured LIRs might be troublesome
to implement - and also thought that newer features, such as 4027 or the
reference counting patch, was a better use of time.
The larger the cache gets, the
Vlad,
You're correct. The existing algorithm is called an Adaptive Replacement
Cache. However, note that this cache does need some proper tuning for
optimal efficiency.
Nicolas
On 2/21/12 12:09 PM, Vladimir Rodionov vrodio...@carrieriq.com wrote:
afaik, existing LruBlockCache is not exactly
If it was ARC (which uses both LRU and LFU) we'd have patenting issues
with IBM, what we have is closer to a 2Q:
http://www.vldb.org/conf/1994/P439.PDF
J-D
On Tue, Feb 21, 2012 at 9:19 AM, Nicolas Spiegelberg
nspiegelb...@fb.com wrote:
Vlad,
You're correct. The existing algorithm is called
I think we should make the BlockCache pluggable for HBase. A simple
reflection-based enhancement to CacheConfig.instantiateBlockCache should do
the trick, is it not? If people think that this is valuable, I can submit a
patch.
This will enable people to play with their own versions of the
In general, I agree about making isolated algorithms pluggable. In this
particular case, I think that Ted was trying to gather consensus on ways
to increase cache efficiency with LIRS being the strawman. I think it's
good to bring up these discussions because it's really easy to add 10k
lines to
On Fri, Feb 17, 2012 at 5:42 PM, Gregory Chanan gcha...@cloudera.comwrote:
Enis,
You mean until Tuesday (the 21st), right?
Ops, my bad. Will be there today.
Enis
Greg
On Fri, Feb 17, 2012 at 5:30 PM, Enis Söztutar enis@gmail.com wrote:
While working on keeping bw compatibility
On Sun, Feb 19, 2012 at 1:45 PM, Mikael Sitruk mikael.sit...@gmail.com wrote:
During compaction the region is not out of service.
According to documentation the max region size for V2 format is 20G
And now the question: Assuming that 20G is the limit and the number of
regions in a single RS
(this is a user question so sending to the relevant mailing list and
putting dev@ in BCC)
Yeah it could be one way. Be aware that some janitorial processes in
the master do a full .META. scan every few minutes you might see very
quick bursts on one machine.
J-D
On Sat, Feb 18, 2012 at 4:57 PM,
This is interesting J.D. so, is there a limitation on the region size or
not? Can it be really any number? If so beside the collection time is there
any impact (perhaps the documentation should be updated too)?
Regarding the number of regions you have (14,398) is it for a single RS?
What is your
On Tue, Feb 21, 2012 at 1:17 PM, Mikael Sitruk mikael.sit...@gmail.com wrote:
This is interesting J.D. so, is there a limitation on the region size or
not?
Your imagination? Like I said nothing blocks you in the code.
Can it be really any number?
That's what it implies.
If so beside the
See inline
On Feb 21, 2012 11:40 PM, Jean-Daniel Cryans jdcry...@apache.org wrote:
On Tue, Feb 21, 2012 at 1:17 PM, Mikael Sitruk mikael.sit...@gmail.com
wrote:
This is interesting J.D. so, is there a limitation on the region size or
not?
Your imagination? Like I said nothing blocks you
On Tue, Feb 21, 2012 at 1:57 PM, Mikael Sitruk mikael.sit...@gmail.com wrote:
If so beside the collection time is there
any impact (perhaps the documentation should be updated too)?
Collection time? You mean GC? Sorry I don't get what you mean.
*Sorry, typo mistake (from mobile) I meant
Ok, so this is approx 150 regions per RS
What are the maths between the memory (index size) and number of regions?
(Btw at the beginning when I mentionned 500 regions it was per RS.)
I'm trying to figure out what should be my cluster configuration, regarding
region, region size, memory size, and
As advertised on the list over the last few weeks, several of us met
today to discuss wire compatibility improvements:
Here's the notes:
http://wiki.apache.org/hadoop/HBaseWireCompatibility20120221
I'll file JIRAs for the subtasks discussed later today.
-Todd
--
Todd Lipcon
Software Engineer,
This describes how they are written, with your knowledge of your data
size and key average size you can do the math:
http://hbase.apache.org/book.html#d0e9542
J-D
On Tue, Feb 21, 2012 at 2:30 PM, Mikael Sitruk mikael.sit...@gmail.com wrote:
Ok, so this is approx 150 regions per RS
What are
See https://builds.apache.org/job/HBase-0.92/296/changes
Changes:
[tedyu] HBASE-5209 Addendum adds znode creation call (David Wang)
--
[...truncated 6668 lines...]
Running org.apache.hadoop.hbase.regionserver.TestMemStore
Tests run: 21, Failures: 0,
On Tue, Feb 21, 2012 at 3:01 PM, Todd Lipcon t...@cloudera.com wrote:
As advertised on the list over the last few weeks, several of us met
today to discuss wire compatibility improvements:
Here's the notes:
http://wiki.apache.org/hadoop/HBaseWireCompatibility20120221
I'll file JIRAs for the
On Tue, Feb 21, 2012 at 1:17 PM, Mikael Sitruk mikael.sit...@gmail.com wrote:
This is interesting J.D. so, is there a limitation on the region size or
not? Can it be really any number? If so beside the collection time is there
any impact (perhaps the documentation should be updated too)?
Yes.
On Tue, Feb 21, 2012 at 12:08 PM, Jean-Daniel Cryans jdcry...@apache.orgwrote:
On Sun, Feb 19, 2012 at 1:45 PM, Mikael Sitruk mikael.sit...@gmail.com
wrote:
During compaction the region is not out of service.
According to documentation the max region size for V2 format is 20G
And now the
In the documentation 20GB is given as an example of a larger size that
can be supported, but nothing blocks you from going way higher than
that. I've done some import tests and had 100GB regions. It just takes
a while to compact the bigger files.
With no impact on Java GC going nuts? FB
On Tue, Feb 21, 2012 at 5:44 PM, M. C. Srivas mcsri...@gmail.com wrote:
With no impact on Java GC going nuts? FB reported (a few months ago) it
was bad to run a region-server
with -Xmx larger than 15G or 16G. Unless its no longer true, wouldn't that
be limiting factor for how
large one
See https://builds.apache.org/job/HBase-0.92/297/
--
[...truncated 6706 lines...]
Running org.apache.hadoop.hbase.rest.TestStatusResource
Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 17.055 sec
Forking command line: /bin/sh -c cd
On Tue, Feb 21, 2012 at 6:16 PM, Stack st...@duboce.net wrote:
On Tue, Feb 21, 2012 at 5:44 PM, M. C. Srivas mcsri...@gmail.com wrote:
With no impact on Java GC going nuts? FB reported (a few months ago) it
was bad to run a region-server
with -Xmx larger than 15G or 16G. Unless its no
On Tue, Feb 21, 2012 at 9:29 PM, M. C. Srivas mcsri...@gmail.com wrote:
Yes, that was my thinking --- to do a major compaction the region-server
would have to load all the flushed files for that region, merge them, and
then write out the new region. If the region-file was 20g in size, the
See https://builds.apache.org/job/HBase-0.92-security/93/changes
Changes:
[tedyu] HBASE-5209 Addendum adds znode creation call (David Wang)
--
[...truncated 6731 lines...]
Running org.apache.hadoop.hbase.rest.TestScannersWithFilters
Tests run: 10,
See https://builds.apache.org/job/HBase-0.92/298/
--
[...truncated 6703 lines...]
Tests run: 14, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.156 sec
Forking command line: /bin/sh -c cd
https://builds.apache.org/job/HBase-0.92/ws/trunk
I've just filed a bunch of JIRAs as subtasks of HBASE-5305 based on
the things we discussed at the meeting:
https://issues.apache.org/jira/browse/HBASE-5305#summary
I didn't fill in any target version yet... before I did so, I had the
idea that we may want to do this work on a feature branch.
On Tue, Feb 21, 2012 at 10:56 PM, Todd Lipcon t...@cloudera.com wrote:
I didn't fill in any target version yet... before I did so, I had the
idea that we may want to do this work on a feature branch.
What do you folks think? It would allow us to commit partial work
without leaving us in a spot
35 matches
Mail list logo