From: Ian Kent <raven@themaw.net>
To: christof.koehler@bccms.uni-bremen.de, autofs@vger.kernel.org
Subject: Re: autofs reverts to IPv4 for multi-homed IPv6 server ?
Date: Sun, 10 Apr 2016 10:17:17 +0800 [thread overview]
Message-ID: <1460254637.3508.12.camel@themaw.net> (raw)
In-Reply-To: <20160408161553.GA15153@bccms.uni-bremen.de>
On Fri, 2016-04-08 at 18:15 +0200, Christof Koehler wrote:
> OK, this actually fails to mount anything at all
That's right, as I say the RPC communication isn't failing in an
unexpected way so we aren't seeing any error messages.
About all that can be done is to add a patch that adds some extra
logging to try and get to the bottom of it.
>
> Apr 8 18:12:53 core324 automount[963]: parse_mount: parse(sun): core
> of entry: options=fstype=nfs4,rw,intr,nosuid,soft,nodev,
> loc=core320:/locals
> Apr 8 18:12:53 core324 automount[963]: sun_mount: parse(sun):
> mounting root /local, mountpoint core320, what core320:/locals, fstype
> nfs4, options rw,intr,nosuid,soft,nodev
> Apr 8 18:12:53 core324 automount[963]: mount_mount: mount(nfs):
> root=/local name=core320 what=core320:/locals, fstype=nfs4,
> options=rw,intr,nosuid,soft,nodev
> Apr 8 18:12:53 core324 automount[963]: mount_mount: mount(nfs): nfs
> options="rw,intr,nosuid,soft,nodev", nobind=0, nosymlink=0, ro=0
> Apr 8 18:12:53 core324 automount[963]: get_nfs_info: called with host
> core320(192.168.220.70) proto 6 version 0x40
> Apr 8 18:12:53 core324 automount[963]: get_nfs_info: called with host
> core320(2001:638:708:1261:2000::70) proto 6 version 0x40
> Apr 8 18:12:53 core324 automount[963]: mount(nfs): no hosts available
> Apr 8 18:12:53 core324 automount[963]: dev_ioctl_send_fail: token =
> 3006
> Apr 8 18:12:53 core324 automount[963]: failed to mount /local/core320
> Apr 8 18:12:53 core324 automount[963]: handle_packet: type = 3
>
> Maybe I am missing dependencies ...
>
>
> On Fri, Apr 08, 2016 at 06:12:26PM +0200, Christof Koehler wrote:
> > Hello,
> >
> > I managed to build a new autofs 5.1.1 from the ubuntu 16.04 source
> > package after throwing out all systemd dependencies. With that
> >
> > root@core324:~/autofs-5.1.1# ldd /usr/lib/x86_64-linux
> > -gnu/autofs/mount_nfs.so
> > linux-vdso.so.1 => (0x00007ffff7ffd000)
> > libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0
> > (0x00007ffff7b73000)
> > libtirpc.so.1 => /lib/x86_64-linux-gnu/libtirpc.so.1
> > (0x00007ffff794b000)
> > libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6
> > (0x00007ffff7585000)
> > /lib64/ld-linux-x86-64.so.2 (0x0000555555554000)
> > libgssglue.so.1 => /lib/x86_64-linux-gnu/libgssglue.so.1
> > (0x00007ffff737b000)
> > libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2
> > (0x00007ffff7177000)
> >
> > However, I had to edit lib/rpc_subs.c as indicated in
> > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=737679, otherwise
> > it would not
> > work (/usr/lib/x86_64-linux-gnu/autofs/mount_nfs.so: undefined
> > symbol: auth_put).
> >
> > With the new binary /local/core330 does not mount either, but the
> > error message is
> > probably an improvement:
> >
> > Apr 8 18:05:25 core324 automount[963]: parse_mount: parse(sun):
> > dequote("core330:/locals"
> > ) -> core330:/locals
> > Apr 8 18:05:25 core324 automount[963]: parse_mount: parse(sun):
> > core of entry: options=fs
> > type=nfs4,rw,intr,nosuid,soft,nodev, loc=core330:/locals
> > Apr 8 18:05:25 core324 automount[963]: sun_mount: parse(sun):
> > mounting root /local, mount
> > point core330, what core330:/locals, fstype nfs4, options
> > rw,intr,nosuid,soft,nodev
> > Apr 8 18:05:25 core324 automount[963]: mount_mount: mount(nfs):
> > root=/local name=core330
> > what=core330:/locals, fstype=nfs4, options=rw,intr,nosuid,soft,nodev
> > Apr 8 18:05:25 core324 automount[963]: mount_mount: mount(nfs): nfs
> > options="rw,intr,nosu
> > id,soft,nodev", nobind=0, nosymlink=0, ro=0
> > Apr 8 18:05:25 core324 automount[963]: get_nfs_info: called with
> > host core330(192.168.220
> > .118) proto 6 version 0x40
> > Apr 8 18:05:25 core324 automount[963]: get_nfs_info: called with
> > host core330(fd5f:852:a27c:1261:2000::118) proto 6 version 0x40
> > Apr 8 18:05:25 core324 automount[963]: get_nfs_info: called with
> > host core330(2001:638:708:1261:2000::118) proto 6 version 0x40
> > Apr 8 18:05:25 core324 automount[963]: mount(nfs): no hosts
> > available
> > Apr 8 18:05:25 core324 automount[963]: dev_ioctl_send_fail: token =
> > 3004
> > Apr 8 18:05:25 core324 automount[963]: failed to mount
> > /local/core330
> > Apr 8 18:05:25 core324 automount[963]: handle_packet: type = 3
> > Apr 8 18:05:25 core324 automount[963]:
> > handle_packet_missing_indirect: token 3005, name core330, request
> > pid 2397
> >
> > I had no success trying to get the 5.0.7 source package to actually
> > link libtirpc,
> > no idea why.
> >
> > Best Regards
> >
> > Christof
> >
> > On Fri, Apr 08, 2016 at 04:29:07PM +0200, Christof Koehler wrote:
> > > Hello,
> > >
> > > apparently I confused my 5.1.1 source built experiment and my
> > > debian
> > > package rebuild experiment when I reported that libtirpc was used
> > > in my
> > > last email. So here is a new try to rebuild the deb source with
> > > --with-libtirpc.
> > >
> > > I did a apt-get source autofs and added --with-libtirpc to debian
> > > rules.
> > > After that it would of course not allow me to build a package,
> > > "aborting
> > > due to unexpected upstream changes". So I just did a "dpkg
> > > -buildpackage
> > > -b" and then dpkg -i autofs... . Attached is the file build.out.gz
> > > which
> > > contains the stdout output. Clearly libtirpc is used somehow in
> > > the build.
> > >
> > > After restoring maps in /etc I did a service restart autofs and
> > > with
> > > debug loglevel I get
> > >
> > > Apr 8 16:20:33 core324 automount[14615]: open_mount:247:
> > > parse(sun):
> > > cannot open mount module nfs
> > > (/usr/lib/x86_64-linux-gnu/autofs/mount_nfs.so: undefined symbol:
> > > clnt_dg_create)
> > >
> > > as reported. I then double checked and actually
> > >
> > > root@core324:~# ldd /usr/lib/x86_64-linux-gnu/autofs/mount_nfs.so
> > > linux-vdso.so.1 => (0x00007ffff7ffd000)
> > > libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6
> > > (0x00007ffff79f3000)
> > > /lib64/ld-linux-x86-64.so.2 (0x0000555555554000)
> > >
> > > no libtirpc.
> > >
> > > I will have to read up on how to properly rebuild the package. The
> > > debian documentation is unfortunately not very user friendly, any
> > > hints are appreciated.
> > >
> > > Best Regards
> > >
> > > Christof
> > >
> > > On Fri, Apr 08, 2016 at 02:25:52PM +0200, Christof Koehler wrote:
> > > > Hello again,
> > > > > I've been thinking about this and I have a couple of thoughts.
> > > > >
> > > > > As far a IPv6 goes using glibc RPC is, I think, not going to
> > > > > work!
> > > > >
> > > > > That's the first thing that needs to be sorted out.
> > > > >
> > > > > I've been using libtirpc in Fedora and RHEL builds for nearly
> > > > > 10 years
> > > > > so I don't think the library problem is with autofs.
> > > > >
> > > > > This is an indication someone is doing something a little
> > > > > dumb:
> > > > >
> > > > > automount[20444]: open_mount:247: parse(sun): cannot open
> > > > > mount module
> > > > > nfs (/usr/lib/x86_64-linux-gnu/autofs/mount_nfs.so: undefined
> > > > > symbol:
> > > > > clnt_dg_create)
> > > >
> > > > concerning my failures to build autofs. First the client has all
> > > > libtirpc packages I think are necessary:
> > > > # dpkg -l libtirpc\*|grep ii
> > > > ii libtirpc-dev 0.2.2
> > > > -5ubuntu2
> > > > ii libtirpc1:amd64 0.2.2
> > > > -5ubuntu2
> > > >
> > > > We have libtirpc1 on the machines by default and I had to
> > > > install libtirpc-dev so that ./configure would conclude that
> > > > --with-libtirpc should do anything.
> > > >
> > > > Actually I tried to compile autofs 5.1.1 from source and a new
> > > > 5.0.7
> > > > package from ubuntu's source deb.
> > > >
> > > > Using the sources at
> > > > https://www.kernel.org/pub/linux/daemons/autofs/v5/
> > > > I was basically confused what to do about the patches. Do I have
> > > > to
> > > > apply everything in patches-5.1.2 to autofs-5.1.1.tar.gz to get
> > > > 5.1.2 ?
> > > > How do I do that automatically ? I noticed that autofs
> > > > -5.1.1.tar.gz
> > > > misses the patch mentioned in message 15 of
> > > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=737679
> > > > but contained in autofs-5.1.1-revert-fix-libtirpc-name
> > > > -clash.patch.
> > > >
> > > > So to make it short I certainly messed something up
> > > > somewhere, the final binary and libs were no success .
> > > > Additionally
> > > > installation did not play nice, although --prefix= was set it
> > > > overwrote
> > > > configuration files in /etc. But I think I
> > > > cleaned everything up afterwards.
> > > >
> > > > If someone can provide some hints I would try it again.
> > > >
> > > > After that I rebuild the 5.0.7 package from source deb after
> > > > adding
> > > > --with-libtirpc to debian/rules as suggested in the bug reports.
> > > > I
> > > > installed from that package. I checked with ldd after
> > > > installing
> > > > that ldd /usr/lib/x86_64-linux-gnu/autofs/mount_nfs.so was build
> > > > with a
> > > > reference to libtirpc.
> > > >
> > > > This try gave the error message in the ubuntu bug
> > > > report
> > > > https://bugs.launchpad.net/ubuntu/+source/autofs/+bug/1564380
> > > >
> > > > So, any hints are appreciated. As long as I can stick to 5.0.7
> > > > rebuilt
> > > > from the source deb installing/re-installing is no problem and I
> > > > can try
> > > > different things you might want. Assuming I can get the program
> > > > to work :-)
> > > >
> > > > Thank you very much for all your help !
> > > >
> > > >
> > > > Best Regards
> > > >
> > > > Christof
> > > >
> > > >
> > > >
> > > > --
> > > > Dr. rer. nat. Christof Köhler email:
> > > > c.koehler@bccms.uni-bremen.de
> > > > Universitaet Bremen/ BCCMS phone: +49-(0)421-218-62334
> > > > Am Fallturm 1/ TAB/ Raum 3.12 fax: +49-(0)421-218-62770
> > > > 28359 Bremen
> > > >
> > > > PGP: http://www.bccms.uni-bremen.de/cms/people/c_koehler/
> > > > --
> > > > To unsubscribe from this list: send the line "unsubscribe
> > > > autofs" in
> > >
> > > --
> > > Dr. rer. nat. Christof Köhler email:
> > > c.koehler@bccms.uni-bremen.de
> > > Universitaet Bremen/ BCCMS phone: +49-(0)421-218-62334
> > > Am Fallturm 1/ TAB/ Raum 3.12 fax: +49-(0)421-218-62770
> > > 28359 Bremen
> > >
> > > PGP: http://www.bccms.uni-bremen.de/cms/people/c_koehler/
> >
> >
> >
> > --
> > Dr. rer. nat. Christof Köhler email:
> > c.koehler@bccms.uni-bremen.de
> > Universitaet Bremen/ BCCMS phone: +49-(0)421-218-62334
> > Am Fallturm 1/ TAB/ Raum 3.12 fax: +49-(0)421-218-62770
> > 28359 Bremen
> >
> > PGP: http://www.bccms.uni-bremen.de/cms/people/c_koehler/
> > --
> > To unsubscribe from this list: send the line "unsubscribe autofs" in
>
--
To unsubscribe from this list: send the line "unsubscribe autofs" in
next prev parent reply other threads:[~2016-04-10 2:17 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-07 14:19 autofs reverts to IPv4 for multi-homed IPv6 server ? Christof Koehler
2016-04-08 4:46 ` Ian Kent
2016-04-08 10:10 ` Ian Kent
2016-04-08 10:14 ` Ian Kent
2016-04-08 12:25 ` Christof Koehler
2016-04-08 14:29 ` Christof Koehler
2016-04-08 15:32 ` Christof Koehler
2016-04-10 2:09 ` Ian Kent
2016-04-08 16:12 ` Christof Koehler
2016-04-08 16:15 ` Christof Koehler
2016-04-10 2:17 ` Ian Kent [this message]
2016-04-10 2:14 ` Ian Kent
2016-04-09 1:42 ` Ian Kent
2016-04-09 9:56 ` Christof Koehler
2016-04-10 2:29 ` Ian Kent
2016-04-25 4:40 ` Ian Kent
2016-04-25 15:06 ` Christof Koehler
2016-04-26 1:06 ` Ian Kent
2016-04-26 9:53 ` Ian Kent
2016-04-26 15:27 ` Christof Koehler
2016-04-27 1:54 ` Ian Kent
2016-04-27 2:27 ` Ian Kent
2016-04-27 16:52 ` Christof Koehler
2016-04-28 2:56 ` Ian Kent
2016-04-28 3:21 ` Ian Kent
2016-04-28 9:12 ` Christof Koehler
2016-04-28 9:10 ` Christof Koehler
2016-04-28 10:50 ` Ian Kent
2016-04-28 11:26 ` Christof Koehler
2016-04-28 12:40 ` Christof Koehler
2016-04-29 1:54 ` Ian Kent
2016-04-29 14:10 ` Christof Koehler
2016-04-29 14:42 ` Christof Koehler
2016-04-30 3:21 ` Ian Kent
2016-04-30 11:36 ` Christof Koehler
2016-04-30 15:15 ` Christof Koehler
2016-04-30 15:16 ` Christof Koehler
2016-05-02 6:01 ` Ian Kent
2016-05-02 16:08 ` Christof Koehler
2016-05-03 7:58 ` Ian Kent
2016-05-03 15:13 ` Christof Koehler
2016-05-04 7:20 ` Ian Kent
2016-05-04 12:38 ` Christof Koehler
2016-04-09 1:35 ` Ian Kent
2016-04-11 2:42 ` Ian Kent
2016-04-11 16:32 ` Christof Koehler
2016-04-11 16:35 ` Christof Koehler
2016-04-12 1:07 ` Ian Kent
2016-04-08 11:47 ` Christof Koehler
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1460254637.3508.12.camel@themaw.net \
--to=raven@themaw.net \
--cc=autofs@vger.kernel.org \
--cc=christof.koehler@bccms.uni-bremen.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.