[dm-devel] [PATCH 3/5] multipath: print warning if multipathd is not running.

[Benjamin Marzinski]

On Thu, Aug 12, 2021 at 10:23:09AM +0000, Martin Wilck wrote:
> On Do, 2021-07-29 at 16:46 -0500, Benjamin Marzinski wrote:
> > If multipath notices that multipath devices exist or were created,
> > and
> > multipathd is not running, it now prints a warning message, so users
> > are
> > notified of the issue.
> > 
> > Signed-off-by: Benjamin Marzinski <bmarzins at redhat.com>
> 
> I'm not sure about this. Are there zero valid use cases for using
> multipath without multipathd?

There certainly are. That's why I tried to limit the warning to only
cases where the user was creating, reloading, or listing devices and
found/created some. In testing, sure, you might want multipathd
disabled, but in general, if you have multipath devices, you want
multipathd running.  If you have it temporarily disabled for some
reason, I don't see much harm in having multipath remind you of that.

> On production systems, I agree,
> multipathd should always be running. Personally wouldn't want to see
> this warning every time I run "multipath" while multipathd is
> (temporarily) disabled. Have you got user requests for this feature?

Well, Support has requested this.  Apparently they have repeatedly seen
cases where people don't have any multipath devices present, because
multipathd is not running.  They run multipath, the devices appear, and
they move on, unaware that multipathd isn't running for whatever
reason.
 
> In particular, I dislike the idea of putting this code into
> libmultipath. I would love to get rid of the "is_daemon" logic some
> day. If at all, the detection of the situation and the warning should
> be implemented in multipath, IMO.

Sure. I can rework this to keep the logic inside multipath,

> The message should be prefixed with the word "Warning: " to make sure
> the admin understands that he's supposed to take action.

Makes sense.

-Ben

> Regards,
> Martin