On Wed, Sep 29, 2010 at 7:09 PM, Pete Zaitcev <zaitcev@xxxxxxxxxx> wrote:
> An interesting question is what to do when iwhd exits. I decided not to
> kill what was started. So, we have a little self-contained cell of
> tabled, chunkd, S3, based off a certain local directory or other
> namespace anchor. Therefore, when iwhd restarts, it tests if the
> cell is still there, and uses that. It also tests if the service
> started successfuly, using the same method.
>
> As we see, for each service iwhd starts, it needs to verify that
> it's available (either before spawning it, or after). This would be
> done best by talking to the service. But iwhd only has S3 client,
> and no CLD client, so it cannot talk to cld (or chunkd). I had an idea:
> add an autostart feature to tabled.
>
> Tabled knows how to talk to both chunkd and cld, so it can verify
> that they are running. It would not be that much code. The downside
> is that it's clearly a special case, encoding of a policy. So I am
> asking how objectionable it would be (including do we want tabled -a
> for tests... they kinda run ok as they are).
It seems like quite a special case feature. tabled is designed to use
multiple chunkd nodes (and hopefully soon, multiple cld nodes). So
having tabled start chunkd/cld seems misaligned with the existing
design.
That said, if it was possible to write a script or program that
performed autostart without modifying tabled, it would be a nice
addition to the git repository. tabled.autostart could be a simple
program that pinged cld/chunkd, started if necessary, then exec'd the
real tabled.
Something modular and separate like that would be great.
Jeff
--
To unsubscribe from this list: send the line "unsubscribe hail-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
