On Mon, Nov 10, 2008 at 12:54 PM, marike <[EMAIL PROTECTED]> wrote:
>
> Thanks for that tip Steve.  Unfortunately, I didn't deploy it with
> capistrano.  But isn't controlling ferret via cap deploy still
> manually controlling it?  I mean every time you want to start|stop
> ferret you still have to do cap deploy as opposed to ./script/
> ferret_server? And would it be possible to re-deploy with capistrano
> now that my site is already running?
>

You can add a cap recipe to bounce ferret and call it in from a deploy
hook. That way both are true. It's automatic and you can bounce it
without redeploying. Here's the thing about ferret, if you don't get
it right then you will have problems and have to rebuild you index and
other downtime things. The only thing I've seen that causes ferret
problems is this scenario:

1.Set up cap to clean up old deploys, like after six.
2. cap deploy. ferret running and everything is good.
3. Deploy 5 more times, everything is OK. (Ferret has not been restarted)
4. Deploy one more time (the seventh deploy) and your app starts
getting ferret crashes.
5. You go crazy and do all kinds of shit, including rebuilding the
index and finally get it fixed.
6. Six deploys later you go through it again.

What's happened is that the DRB ferret is pointing to the symlinked
index dir. Until cap cleans up that original release dir. The real
index dir is fine, but the symlink that the drb server is using is
pointing to the now non-existent release dir.

The trick is to bounce ferret before the sixth deploy to get a new
symlink. Or much easier to just bounce it with every deploy. If this
is done then ferret is rock solid (in my opinion). I haven't had any
need to bounce ferret outside a deploy, once I figured all this out.

Does this make sense? I'm gonna write this up as a blog post.



>
> Lastly, I am still puzzled by what ./script/ferret_server returns.
> For example,
>
> Mac:jazinc marike$ ./script/ferret_server -e production start
> starting ferret server...
> Loaded suite ./script/ferret_server
> Started
>
> Finished in 0.00056 seconds.
>
> 0 tests, 0 assertions, 0 failures, 0 errors
>
> Sazima mentioned that this is normal output, but the 'Finished in
> 0.00056 seconds' implies that aaf/ferret quits. Is there specific
> output in ferret_index.log that will clue me in as to whether or not
> ferret is behaving normally?
>

This odd output is because one of the plugins is requiring
'test_helper' or some thing in test so when init is called it outputs
this stuff. It outputs it when you restart your app too. The thing to
do is to find out which plugin and delete that one require. If you
find it, let me know and I'll delete it in master.

cheers,
steve

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google
Groups "Lovd by Less" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/lovdbyless?hl=en
Who loves ya baby?
-~----------~----~----~----~------~----~------~--~---

Reply via email to