[distcc] distcc problems...

[Martin Pool]

On 10 Mar 2004, Daniel Kegel <dank at kegel.com> wrote:
> ahasty at chiaro.com wrote:
> >However, when I have multiple builds going at the same time, using the 
> >same set of servers, I get one of the following three messages.
> >
> >distccmon-text[10782] (dcc_close) ERROR: failed to close fd4: Stale NFS 
> >file handle
> >or
> >distccmon-text[960] (dcc_mon_read_state) Warning: wrong magic number: 
> >/home/ahasty/.distcc/state/binstate_12563
> 
> It seems you're being bitten by a feature of distcc that you probably aren't
> using: the state files that let other programs watch distcc's state.

If he's running distccmon-text then he is attempting to make use of
them.  But perhaps he can live without it.

> Am I right in assuming these files are not used by distcc itself?

They are written but not read.  But a similar technique is used for
lock files, which are not optional.  If your filesystem is flaking
out, there may be more problems than just the monitor failing.  (In
fact, there have been one or two previous reports of build failures
caused by NFS problems.)

> It seems that this feature can cause problems; perhaps it should be
> off by default.  Thanks for the warning, I might well have run into this
> myself soon.

Perhaps the bug should be fixed instead.

> You can probably work around this by setting DISTCC_DIR to some
> directory on a local hard drive that you have write permission to.

That might be a good idea.  tmpfs (or whatever it's called on Solaris)
might work too.

> Or perhaps you could just gut dcc_note_state() so it becomes a no-op
> if you don't use any of the monitors.

I don't think that's a good idea; it might just hide the problem.
Let's work out what is really wrong.

-- 
Martin 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.samba.org/archive/distcc/attachments/20040311/932553c4/attachment.bin