one namenode partition went down. The right way to reconnect?

[Ayon Sinha <ay...@yahoo.com>]

So we had an NFS mount and a local disk for dfs.name.dir. Our NFS mount lost connection a week ago. So the question is that if we now remount it and it has data from back then, will the namenode detect the staleness and bring it in sync with the other partition? Or should we manually delete the files from that partition and and remount (by the time we copy and remount it might again be stale for quite a few edits)?
 
-Ayon
See My Photos on Flickr
Also check out my Blog for answers to commonly asked questions.

Re: one namenode partition went down. The right way to reconnect?

[Harsh J <ha...@cloudera.com>]

Ayon,

I believe these parts have undergone some changes (and some fixes)
over time so the answer may vary. What version or distro of Hadoop are
you running? Ideally, a simple restart is sufficient, since the NN
will look and identify the most recent name dir based on the edits
timestamps.

Running on CDH3's sources here tells me that just restarting would do.
If I had dfs.name.dir.restore set to true, I wouldn't even require
doing that (it would auto-check at every checkpoint).

On Fri, Aug 26, 2011 at 4:47 AM, Ayon Sinha <ay...@yahoo.com> wrote:
> So we had an NFS mount and a local disk for dfs.name.dir. Our NFS mount lost
> connection a week ago. So the question is that if we now remount it and it
> has data from back then, will the namenode detect the staleness and bring it
> in sync with the other partition? Or should we manually delete the files
> from that partition and and remount (by the time we copy and remount it
> might again be stale for quite a few edits)?
>
> -Ayon
> See My Photos on Flickr
> Also check out my Blog for answers to commonly asked questions.
>



-- 
Harsh J