zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vitalii Tymchyshyn <...@tym.im>
Subject Re: Partitioned Zookeeper
Date Mon, 26 May 2014 08:46:34 GMT
On second thought some kind of wayback machine is needed for the cold
recovery/coordinator change for the case when transaction N is applied and
N-1 is not. But it should be enough to have kind of small buffer for
transactions that are "on the fly".
Note that transaction is confirmed to the client only when all previous are
confirmed. So it would be really small buffer.

Best regards, Vitalii Tymchyshyn
26 трав. 2014 11:32, користувач "Vitalii Tymchyshyn" <vit@tym.im> написав:

> My 5 cents.
> As far as I understand, there are two narrow points in Zookeeper:
> 1) single-threaded transaction numbering on coordinator
> 2) writing data.
> Number 1 IMHO can't be avoided, but it can potentially be very fast as
> single AtomicLong is needed.
> After transaction numbering is done, all other work, including data
> writing, can be distributed and parallelized as any recovery / reordering
> can use this sequential numbering to ensure that there are no gaps in
> transactions sent to clients.
> And actually, I'd use some hashing for workload distribution to ensure
> that there are no hot spots.
> Are there any weak points in my logic?
> Best regards, Vitalii Tymchyshyn.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message