Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Zookeeper >> mail # user >> Prune Txn Log different dataLog and dataDir

Copy link to this message
Re: Prune Txn Log different dataLog and dataDir
On Fri, Oct 26, 2012 at 6:50 AM, Roshan Punnoose <[EMAIL PROTECTED]> wrote:
> Yeah that makes sense. I tried both the zkCleanup.sh that is packaged with Cloudera and the PruneTxnLog that is recommended by the Zookeeper Administrator guide. It is probably something that is fixed in later releases of Zookeeper?

afaik that code hasn't changed in quite a while. I'm not aware of any
outstanding bugs, etc...

Try adding "set -x" to the top of the zkCleanup script and see what
parameters it's using? Ensure it's what you expect - notice that it's
running the prune command twice, once on the datadir and once on the
datalogdir. (or should be at least)

Also, why do you think the log dir is not pruned? Too many files left
behind? prune doesn't remove everything, depending on the parameters
it keeps around a few files. Perhaps you could attach a directory
listing (ls -la) for both dirs?


> On Friday, October 26, 2012 at 9:44 AM, Ted Dunning wrote:
>> You don't have Cloudera Zookeeper. You may have Apache Zookeeper that is
>> packaged with Cloudera's software distribution.
>> On Fri, Oct 26, 2012 at 8:31 AM, Roshan Punnoose <[EMAIL PROTECTED] (mailto:[EMAIL PROTECTED])>wrote:
>> > I have my cloud era zookeeper (3.3.5) running with the dataLogDir and the
>> > dataDir going to two different directories. I've tried many combinations of
>> > the PruneTxnLog command and only the dataDir seems to be cleaned. For some
>> > reason the dataLogDir is not being cleaned as well. Is there something I am
>> > missing?
>> >
>> > --
>> > Roshan Punnoose
>> > Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
>> >