zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Patrick Hunt <ph...@apache.org>
Subject Re: ZK fsync warning
Date Tue, 27 Nov 2012 17:42:54 GMT
On Tue, Nov 27, 2012 at 4:40 AM, Brian Tarbox <briantarbox@gmail.com> wrote:
> Does this warning mean a: the log is too big, b: zk is competing with another process
for disk access, c: other?

That's timing how long the fsync takes on the txnlog, typically it
runs log because the disk is busy, or the OS has a large number of
dirty pages on the volume (ext3 esp), etc...

> How should one respond to the warning?

The admin and troubleshooting guides both provide some insight, but in
particular the txnlog should be on a dedicated spindle (non-vm obv) if
you care about performance (latency). Any fsync timing that's large
than/near your session timeout is going to be trouble.


> On Nov 27, 2012, at 2:47 AM, Patrick Hunt <phunt@apache.org> wrote:
>> On Mon, Nov 26, 2012 at 11:43 PM, Narayanan A R
>> <narayanan.arunachalam@gmail.com> wrote:
>>> Have you seen this before?
>>> 2012-11-25 16:01:08,074 [myid:1] - WARN  [SyncThread:1:FileTxnLog@321] -
>>> fsync-ing the write ahead log in SyncThread:1 took 1470ms which will
>>> adversely effect operation latency. See the ZooKeeper troubleshooting guide
>>> I am running ZK cluster of size 3 in a VM.
>> Yes, I added it recently to highlight situations where environment
>> could adversely effect ZK latency.
>> Patrick

View raw message