Re: LMDB killed process and LOCK_MUTEX_W()

[Howard Chu <hyc@symas.com>]

Dimitrios Apostolou wrote:
> Hi Howard,
>
> On Wed, 16 Jul 2014, Dimitrios Apostolou wrote:
>
> > On Wed, 16 Jul 2014, Howard Chu wrote:
> >
> > > Dimitrios Apostolou wrote:
> > > > What would you suggest for such a situation? I'm thinking of patching LMDB
> > > > to lock with mutex_timedwait() and periodically check if the PID having
> > > > taken the mutex is still alive. Is the writer PID stored somewhere, or a
> > > > change of format will be needed? Any other ideas are welcome!
> > >
> > > We have a patch to use robust mutexes. They're a few percent slower but
> > > will allow recovery from this situation.
> >
> > Great! Is the patch public so that I can see if it fixes the situation?
>
> By using Google I could find a patch from 2012 from Hallvard Furuseth at
> http://folk.uio.no/hbf/OpenLDAP/its7165.diff is that the one you are
> talking about?

Not quite, that's a much older patch and it primarily contains code to break 
threading.
> In addition, is the PID of the lock-acquirer stored somewhere in shared
> map? There are lots of PIDs and TIDs in there and I'm having trouble in
> deciphering the code, so I'd appreciate if you pointed to the struct
> member.
>
>
> Thanks!
> Dimitris


--
  -- Howard Chu
  CTO, Symas Corp.           http://www.symas.com
  Director, Highland Sun     http://highlandsun.com/hyc/
  Chief Architect, OpenLDAP  http://www.openldap.org/project/