Date: Sun, 23 Jan 2022 16:10:40 +1030 From: "Daniel O'Connor" <darius@dons.net.au> To: tech-lists <tech-lists@zyxst.net> Cc: freebsd-arm@freebsd.org, freebsd-current@freebsd.org Subject: Re: POLLHUP detected on devd socket Message-ID: <8BA9DE0C-81B4-420E-B326-7B7133EC31BA@dons.net.au> In-Reply-To: <YewV/8v7nb78a%2Bg4@ceres.zyxst.net> References: <YewV/8v7nb78a%2Bg4@ceres.zyxst.net>
index | next in thread | previous in thread | raw e-mail
> On 23 Jan 2022, at 01:04, tech-lists <tech-lists@zyxst.net> wrote: > context is a usb3-booting rpi4 running main-n252544-7406ec4ea99 > > I see this in /var/log/daemon.log: > > Jan 22 13:24:54 redacted zfsd[886]: POLLHUP detected on devd socket. > Jan 22 13:24:54 redacted zfsd[886]: Disconnecting from devd. > Jan 22 13:24:54 redacted zfsd[886]: ConnectToDevd: Connecting to devd. > Jan 22 13:24:54 redacted zfsd[886]: Unable to connect to devd > Jan 22 13:24:54 redacted zfsd[886]: Disconnecting from devd. > > When this happens, several other things happen: > > 1. the sshd server dies. Already-established sessions still run, but the server itself stops and no new ssh connections can be made. > > 2. syslogd dies. so does exim, nginx, smartd, sshguard, pf > > I don't know where to start looking for reasons for this. Is it > a known issue? In the meantime I've turned off zfsd as it's not > really needed in this machine's context (headless, zfs on single > disk), from what I can tell. Is it reproducible? ie can you trigger it post boot? It is very strange that devd dying would kill anything else running - I have stopped and started it a lot while testing devd scripts and I haven't seen anything else break. One idea would be to hack up devd's rc.d script and run it under ktrace and try and get some clues. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaumhelp
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8BA9DE0C-81B4-420E-B326-7B7133EC31BA>