hadoop-zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sam Baskinger <sam.baskin...@networkedinsights.com>
Subject Re: Demo Code: Shared/Exclusive Lock
Date Wed, 05 May 2010 20:54:56 GMT
Mahadev and Henry,

Thanks for the warm welcome and patience. :)

The code is lightly used and has been manually tested to block and release. As Networked Insights
transitions code out of its labs it generally matures the code through a refactor and adding
additional tests. This code is still somewhat young in the process.

Regarding what it does differently / adds, if I'm reading the existing code correctly it adds
the concept of a shared lock where many processes may lock until the exclusive lock is reached.
Only one process may hold the exclusive lock. The use case we had in mind was kicking off
a herd of machines to perform write operations (exclusive lock) on HDFS data, but not if there
was a process reading it (shared lock).

I would love this sort of artifact to exist in the core of ZooKeeper, but that's another email
thread. :)

I think everyone has see the JIRA issue pop up (ZOOKEEPER-767). Patch is attached.

Let me know what I can do to make this process easier! Again, thanks for the warm reception.

Sam Baskinger
Software Engineer
Networked Insights

On May 5, 2010, at 3:18 PM, Mahadev Konar wrote:

Great to see this!!!!

IT would be good to have this contributed. Sam, how does this compare to
locks implementation currently in src/recipes? Anyhow a well tested and used
implementation would nevertheless be useful!

You can open a jira and submit your code to it. I am pretty committers will
be willing to work with you to get this in the codebase!

Thanks
mahadev


On 5/5/10 1:06 PM, "Sam Baskinger" <sam.baskinger@networkedinsights.com<mailto:sam.baskinger@networkedinsights.com>>
wrote:



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