Does your regions have store files in them?
In HBase all the store files are opened and the readers are kept open
always when the region opening happens. When there is a compaction
happening the count of those file handlers open gets reduced based on the
number of files compacted.
Regards
Ram
O
Dear All,
Observation:
I have HBase cluster with Kerberos enabled, when the Region Server startups
then we observed some 20-30 socket fds are in CLOSE_WAIT state,
After that when the Region Server starts opening then the socket fds in
CLOSE_WAIT starts increasing gradually (almost same as number
You can extend RowFilter by extracting the reverse timestamp suffix and
compare it with the time range.
In filterRowKey() method, if the suffix falls within the time range, return
false. Otherwise return true.
Cheers
On Wed, Jul 23, 2014 at 12:18 PM, Mingtao Zhang
wrote:
> Thank you :) ... sor
Please take a look at:
HBASE-9360 Enable 0.94 -> 0.96 replication to minimize upgrade down time
Cheers
On Wed, Jul 23, 2014 at 10:12 AM, Guangle Fan wrote:
> Not as described in the instruction page shutting down the cluster and do
> in-place upgrading. Yes, the process is to migrate table to
When this happens again, please capture jstack and pastebin it so that we
can know whether the deadlock was due to synchronization on masterAndZKLock.
If possible, please upgrade to 0.98.4
Cheers
On Wed, Jul 23, 2014 at 6:41 AM, Ted Yu wrote:
> There was no line number shown in the image.
>
>
Thank you :) ... sorry, seems I didn't make the question clear :(.
My rowkey looks similar to this
murmurHashing() + 'reversed time stamp'
I am looking for a way to say 'get all rows within a time range'.
Best Regards,
Mingtao
On Wed, Jul 23, 2014 at 2:43 PM, Jean-Marc Spaggiari <
jean-m...@s
Hi Mingtao,
You can use Fuzzy filter for this, or regular expression.
You can take a look here:
https://hbase.apache.org/apidocs/org/apache/hadoop/hbase/filter/FuzzyRowFilter.html
Fuzzy might not be able to match your last byte variation. You might need
the Regex filter for that:
RowFilter rowFi
Thank you all. Moved to Murmur hash.
Best Regards,
Mingtao
On Mon, Jul 21, 2014 at 10:58 PM, Ishan Chhabra
wrote:
> No *guarantees* on collision, but yes, it is a deterministic mapping and
> you won't see collisions in that range (provided you choose enough bits).
>
> See MurmurHash here: http
Hi All,
Is it possible to pad the first byte while specifiy the following bytes for
rowkey filter?
19890208~19890209 will select from the following if
availble:
19890208
000119890208
000219890208
...
19890208
19890209
...
19890209
Thank yo
Not as described in the instruction page shutting down the cluster and do
in-place upgrading. Yes, the process is to migrate table to another cluster
without service downtime. The other cluster is already 0.96. How is the
process of doing migration ?
On Tue, Jul 22, 2014 at 7:34 PM, Ted Yu wrote
There was no line number shown in the image.
HBASE-11460 was fixed in 0.98.4 which got released yesterday. There is some
similarity between stack trace shown in your image and the second stack trace
of HBASE-11460.
Cheers
On Jul 23, 2014, at 5:24 AM, Anand Nalya wrote:
> The image can be f
The image can be found at http://i.imgur.com/JTXw1cg.png
- Anand
On 23 July 2014 18:44, Ted Yu wrote:
> Your image attachment didn't go through.
>
> Mind posting it on third party website and provide a link ?
>
> Cheers
>
> On Jul 23, 2014, at 5:08 AM, Anand Nalya wrote:
>
> > Hi,
> >
> > I'm
Your image attachment didn't go through.
Mind posting it on third party website and provide a link ?
Cheers
On Jul 23, 2014, at 5:08 AM, Anand Nalya wrote:
> Hi,
>
> I'm using HBase 0.98.3 with hadoop 2.4.0. We have application that calls
> HTable.put and in case of exception we have we cre
Hi,
I'm using HBase 0.98.3 with hadoop 2.4.0. We have application that calls
HTable.put and in case of exception we have we create the table.
The application is freezing in unexpected manner. YourKit is
showing HConnectionManager.HConnectionImplementation#locateRegionInMeta as
frozen. Please find
Hi all,
Thanks for reply.
Finally it worked by following post
http://www.spaggiari.org/index.php/hbase/how-to-install-snappy-with-1#.U8Pui-9ZuZY
.
Issue was:
By default hadoop-2 comes with 32 bit native library and without native
support so i need to compile source code of hadoop for 64 bit ar
Logs here, all periodicFlusher and cacheFlusher, I don't see any other logs.
2014-07-23 18:20:37,227 INFO [regionserver60020.periodicFlusher]
regionserver.HRegionServer: regionserver60020.periodicFlusher requesting flush
for region contacts,5c,1405328195955.804e9746463e04db5184c898e8a2324e. a
If flush interval is set to 1 hr, that wouldn't explain why you frequently
encountered slow performance for deletes, right ?
Mind pastebin'ing portion of region server log ?
Cheers
On Jul 23, 2014, at 1:57 AM, "Li Yong" wrote:
> Thanks, Ted.
> I see thousands upon thousands 'periodicFlush
Thanks, Ted.
I see thousands upon thousands 'periodicFlusher requesting flush for
region' logs. I guess the conf 'hbase.regionserver.optionalcacheflushinterval'
effects this! We use the default 1h.
Is this the reason delete delay too much? And how can I repair this
without shutdown
Have you checked region server logs to see if there was some clue ?
If so, mind pastebin'ing some snippet ?
Cheers
On Jul 23, 2014, at 12:22 AM, "Li Yong" wrote:
> Hi all,
> I move my cluster to another cluster with CopyTable, then I found too
> many timeout, Most of them are Delete! I fo
Hi all,
I move my cluster to another cluster with CopyTable, then I found too many
timeout, Most of them are Delete! I found the table has too many storefiels, so
i do a major_compact, then things be better. After an hour, timeout again! I
don't know why so many delay in delete, not put, no
20 matches
Mail list logo
