hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jean-Marc Spaggiari <jean-m...@spaggiari.org>
Subject Re: Put while move or split
Date Mon, 03 Dec 2012 17:30:39 GMT
Thanks all for your comments.

I was talking about the 3 minutes period because by default, ZK will
have a 3 minute timeout. So if a server crash, it can take up to 3
minutes to Zookeeper to detect that. And in this meantime, the client
might not be able to access the impacted regions.

So what I was wondering is, what's the right wait to handle that.
Should we always catch all the failures of all the put/get and others
and have a backup plan? Or is sometime hbase taking care of that.

So what I have understand here is that HBase will take care of the
put/get when regions are "simply" moving and temporary unavailable
(compaction, etc.) but the client should take care of bigger issues
like when the region is not available for a longer period for any


2012/12/3, Harsh J <harsh@cloudera.com>:
> Hey Mike,
> A move is logical in terms of a region and region servers (i.e. only
> an ID and a command is passed around), but yes the location of blocks
> could impact its performance a bit (until a further compaction), but
> would not prevent availability.
> On Mon, Dec 3, 2012 at 3:53 PM, Michel Segel <michael_segel@hotmail.com>
> wrote:
>> He could have meant that the region was unavailable for a longer
>> duration.
>> My take is that you could have limited bandwidth, and a very large region
>> on the move.
>> While I don't imagine that it could take that long to move a region, you
>> never know how things are tuned and you could in theory hit a timeout.
>> I also have to wonder if HBase relies on the underlying HDFS to handle the
>> move.
>> Thoughts?
>> Sent from a remote device. Please excuse any typos...
>> Mike Segel
>> On Dec 3, 2012, at 3:45 AM, ramkrishna vasudevan
>> <ramkrishna.s.vasudevan@gmail.com> wrote:
>>> Hi Jean
>>> Incase of Puts and Scans if the split is in progress it will retry.
>>> Suppose before the client gets to know about the split, still if the put
>>> is directed with the parent region row key, internally HBase directs it
>>> to
>>> the correct daughter region.
>>> But i am not clear in what you mean by Region server down for 3 mins?
>>> You
>>> mean during split your RS went down?
>>> Regards
>>> Ram
>>> On Mon, Dec 3, 2012 at 8:23 AM, Jean-Marc Spaggiari
>>> <jean-marc@spaggiari.org
>>>> wrote:
>>>> Ok. so if on the client side I'm "simply" using Puts, scans and gets
>>>> fron the HTableInterface, I don't have to worry about that since it
>>>> will retry.
>>>> But I still should consider catching exceptions when regions are not
>>>> available because the region server is down for less than 3 minutes,
>>>> right?
>>>> Thanks,
>>>> JM
>>>> 2012/12/2, Nicolas Liochon <nkeywal@gmail.com>:
>>>>> It's not accessible, but it's more or less transparent (latency impact
>>>>> aside) for the end user: the hbase client will retry the operation.
>>>>> On Sun, Dec 2, 2012 at 11:10 PM, Jean-Marc Spaggiari <
>>>>> jean-marc@spaggiari.org> wrote:
>>>>>> ?
> --
> Harsh J

View raw message