FTR, the hosts where I had this problem now run Jessie & systemd. nfs-utils 1.2.x under systemd is completely broken; I had to write my own NFS units from scratch (inc. rpcbind, IIRC).
The end result is: my current bugs are new, and I don't care about this old one. Your analysis that the real cause is probably #804670 sounds sensible. (PS: FTR, nfs-utils 1.3 should Just Work under systemd, but I haven't gotten around to testing it yet.) > Date: Thu, 26 Oct 2017 19:01:38 +0200 > From: Laurent Bigonville <bi...@debian.org> > To: 706691-cl...@bugs.debian.org > Subject: Re: In rc0.d, sendsigs stops before rpcbind stops > Message-ID: <5eaf5b96-9f76-ca62-9951-a1bafa406...@debian.org> > User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 > Thunderbird/52.4.0 > > On Fri, 3 May 2013 22:07:19 +1000 "Trent W. Buck" <trentb...@gmail.com> > wrote: > > > In a minimal live-boot image I built, I noticed that sendsigs was > > running before all the NFS stuff was turned off. > > I know that this bug is really old, but there are something that I don't > understand. > > The complete point of the "sendsigs.omit" is (was?) to avoid killing > processes when the sendsigs is stopped. > > The fact that rpcbind is still running after sendsigs IS expected and > rpcbind is stopped just after all the NFS file systems have been unmounted > and the nfs-daemon are stopped. > > For me this is expected behavior. It's however possible that due to #804670, > rpcbind is not killed when the **rpcbind** initscript returns, that will be > fixed > > Feel free to reopen if the I misunderstood something.
