Hello,

I have writen a quick'n'dirty init script for EnhanceIO.

On start, it checks if EnhanceIO may have been active already by looking for its udev script.
If not, if sources its configuration from /etc/default/enhanceio and brings up the cache.

On stop, it switches the cache to read-only and blocks till all changes have been written to disk.
Afterwards it deletes the cache.

Cheers,
Unki

Hi Andreas,

Thank you but I doubt that it would be a good idea to manage caching by init
script. I'm also concerned about maintenance effort and possible situations
when this approach may fail.

I'm convinced that such improvements should be introduced upstream.
Unfortunately upstream is not very active... Can you open the issue (or pull
request) upstream?

Have you seen

https://github.com/stec-inc/EnhanceIO/pull/54

What do you think about it? Wouldn't improving existing udev handler would be
better? Can you try and review it?

Besides I'd like to discourage people from using write-back caches as much as
I can...

For a moment I'll mark this "wontfix" due to lack of upstream feedback.
Also I couldn't find time to test init script ever since it was reported...

Finally I'd like to note that "/dev/sd[a-z]" notation is not persistent hence
it is not reliable to use on systems with multiple HDDs.

Thank you.

---

The Santa myth is one of the most effective means ever devised for
intimidating children, eroding their self-esteem, twisting their
behavior, warping their values, and slowing their development of
critical thinking skills.
        -- Tom Flynn, "The Trouble with Christmas"