Re: HBase and HDFS replication

Take a look at the zookeeper session timeout. The ephemeral node of the rs going down will be deleted when session is expired and then other regionservers will race to take ownership of the regions being down. The default session timeout is too high so I think it may be related to the problem you a

Re: Replication metrics with more than one Peer/Slave

it was fixed in 0.95 > > On Thursday, November 28, 2013, Pablo Medina wrote: > > > Hi all, > > > > Knowing that replication metrics are global at the region server level in > > HBase 0.94.13, what is the meaning of a metric like sizeOfLogQueue when > > repl

Replication metrics with more than one Peer/Slave

Hi all, Knowing that replication metrics are global at the region server level in HBase 0.94.13, what is the meaning of a metric like sizeOfLogQueue when replicating to more than one peer/slave? Is it the queue size reported by the "last" replication source thread ? does the last thread win ? Can

Re: Performance penalty: Custom Filter names serialization

> On Tue, Aug 20, 2013 at 8:56 AM, Pablo Medina >wrote: > > > Hi all, > > > > I'm using custom filters to retrieve filtered data from HBase using the > > native api. I noticed that the class full names of those custom filters > is > > being sent as the

Performance penalty: Custom Filter names serialization

Hi all, I'm using custom filters to retrieve filtered data from HBase using the native api. I noticed that the class full names of those custom filters is being sent as the bytes representation of the string using Text.writeString(). This consumes a lot of network bandwidth in my case due to using

Connection pool

Hi all, What is the recommended strategy/configuration regarding connection pooling in production? I have read the HBase definitive guide section about it and some old threads in the mailing list which suggests to use an HTablePool sharing a unique HConnection. I'm wondering if this can be a bottl

Re: How many column families in one table ?

Lars, when you say 'when one memstore needs to be flushed all other column families are flushed', are you referring to other column families of the same table, right? 2013/8/4 Rohit Kelkar > Regarding slow scan- only fetch the columns /qualifiers that you need. It > may be that you are fetch

Re: help on key design

to 4 bytes and prepending it to his key. This should give > him some randomness. > > > > On Jul 31, 2013, at 1:57 PM, Pablo Medina wrote: > > > If you split that one hot region and then move a half to another region > > server then you will move the half of the loa

Re: help on key design

on key design > >> > >> Thanks for the responses! > >> > >>> why don't you use a scan > >> I'll try that and compare it. > >> > >>> How much memory do you have for your region servers? Have you enabled > >>> bloc

Re: help on key design

x27;ll try that and compare it. > > > > > How much memory do you have for your region servers? Have you enabled > > > block caching? Is your CPU spiking on your region servers? > > Block caching is enabled. Cpu and memory dont seem to be a problem. > > > >

Re: help on key design

The scan can be an option if the cost of scanning undesired cells and discarding them trough filters is better than accessing those keys individually. I would say that as the number of 'undesired' cells decreases the scan overall performance/efficiency gets increased. It all depends on how the keys

Re: Handlers being blocked during reads

I've just created a Jira to discuss this issue: https://issues.apache.org/jira/browse/HBASE-9087 Thanks! 2013/7/30 Elliott Clark > On Mon, Jul 29, 2013 at 11:08 PM, lars hofhansl wrote: > > Do you think we should change it to use a ConcurrentHashMap > > Yea, I think that would be great. I re

Re: Handlers being blocked during reads

I got that stack trace using jstack Thanks, Pablo. 2013/7/30 hua beatls > how to get the 'stack trace'? > > Thanks! > > > beatls > > > On Tue, Jul 30, 2013 at 11:20 AM, Pablo Medina >wrote: > > > Hi all, > > > > I'm ha

Re: Handlers being blocked during reads

gt; What do you see in the logs around the time you saw that behavior? is > this happening on a single Region Server? what version of HBase are > you running? > > cheers, > Esteban. > > Cloudera, Inc. > > On Jul 29, 2013, at 20:21, Pablo Medina wrote: > > > Hi

Handlers being blocked during reads

Hi all, I'm having a lot of handlers (90 - 300 aprox) being blocked when reading rows. They are blocked during changedReaderObserver registration. Does anybody else run into the same issue? Stack trace: "IPC Server handler 99 on 60020" daemon prio=10 tid=0x41c84000 nid=0x2244 waiting on

Handlers being blocked during reads

Hi all, I'm having a lot of handlers (90 - 300 aprox) being blocked when reading rows. They are blocked during changedReaderObserver registration. Does anybody else run into the same issue? Stack trace: "IPC Server handler 99 on 60020" daemon prio=10 tid=0x41c84000 nid=0x2244 waiting on