Your question implies a lot of other considerations so I'll answer
strictly to it. By default we retry 10 times to get access to data
with pauses of increments of 2 seconds in the same fashion of the BSD
TCP syn backoff table.
The configs are:
hbase.client.pause
2000
General client
Thanks JD for the detailed reply.
Does the underlying java api currently block in case if region is not available
? I would like to get an immediate retry indication for the java call in such
cases so that I can redirect the request to the duplicate table in the other
data center. Can this be s
This would be a good FAQ entry Andrew, your explanations are way
better than mine. The master publishing itself is new in 0.20 and I
guess others will hit this issue with Ubuntu.
J-D
On Tue, Nov 24, 2009 at 2:03 PM, Andrew Purtell wrote:
> Good to hear Mark.
>
> HBase is sensitive to this in a w
> I have a hbase table which created via mapred tool, took almost 1 hour to
> load by the loadtable.rb script and to be available for serving. The scale
> was 8k regions per server. I am on 0.20.1, r822817 though. I am yet to test
> the failure case, but it will take around 1hour/ no.of RS to
That error means the Namenode isn't giving the locations for a file in
the namespace, I don't think this is the error you are searching for.
Which tests failed? Which version? Is it because of errors or because
of failed assertions? Look at the end of the TEST-* files to figure
it.
J-D
On Wed, N
First, there is work under way for 0.21 which will shorten the time necessary
for region redeployment. Part of the delay in 0.20 is less than ideal
performance in that regard by the master.
Beyond that, just as a general operational principle, I recommend that you host
no more than 200-250 reg
Forgive what might be a dumb question, but did you change your HBase rootdir to
something other than the default (which is in /tmp)?
In the HBase shell there is a command 'flush' which will force persistence of
anything in memstore, but this should not be needed in normal operation unless
you
Try this: Do not cache on the crawlers, just write through.
Run each region server with plenty of heap (4 GB to start). So it seems you
need more RAM on your systems, or you should move your crawlers off to separate
servers to free up RAM and CPU.
Adjust your HBase site config as follows:
Hi, I'm Patrick from the ZooKeeper team. Recently we've (the ZK dev
community) been working more closely with the HBase dev team. In
particular to ensure that ZK-HBase interaction is the best it can be and
improve things where it's not. We are also looking at how HBase might
take more advantage
I'm not sure I understand your requirements entirely, but there are order
insensitive functions you can use to generate a key for any pair of
elements. Any commutative operation on the two keys would work, but if you
use a hash you need to worry about collisions. If you take your two (or
more) ke
I just ran the "ant test" and a bunch of the tests failed. Upon looking at the
logs, this one error came up:
Could not get block locations. Source file "/user/hadoop/hbase.version" -
Aborting...
java.io.IOException: Could not get block locations. Source file
"/user/hadoop/hbase.version" - Abort
Hi,
I was just wondering whether Linux HA (HeartBeat + DRDB) could be an
option in this case. What you guys think?
- Imran
On Wed, Nov 25, 2009 at 6:29 PM, Murali Krishna. P
wrote:
> Hi Ryan,
> Thanks for the quick response.
>
> We are planing have this in 2 or 3 data centers for BCP and la
Hi Ryan,
Thanks for the quick response.
We are planing have this in 2 or 3 data centers for BCP and latency
reasons. Currently application runs in a non-scalable cluster, essentially we
have the data partitioned across multiple fixed columns. The entire cluster of
machines can be conside
With multiple masters, the election is mediated by zookeeper and the
idle masters are awaiting the relection cycle.
The problems with brining regions up after a failure isnt the actual
speed of loading them, but bugs with the master. This is being fixed
in 0.21. It will allow us to much more rapi
14 matches
Mail list logo
