View previous topic :: View next topic |
Author |
Message |
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 4:45 am Post subject: HELP: Can't fix mount.nfs: Stale file handle error? [SOLVED] |
|
|
Wow wow wow. I recall getting this years ago, but not in many years...though I don't ever recall it being this bad: I use an NFS mount between my mythtv frontend to a video directory on the backend system. Today I noticed it wasn't mounted...no idea why. Trying to mount I'm getting:
Code: | mount /mnt/remote_media
mount.nfs: Stale file handle |
Even after rebooting BOTH machines. I've read more crap than I care to talk about on this and nothing's working. Any suggestion would be welcome.
Tom
Last edited by tld on Sat Sep 28, 2019 1:57 pm; edited 1 time in total |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 4:51 am Post subject: |
|
|
I just tried from another machine and I get the same thing, so clearly it's something on the server. Wow...horrible.
Tom |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 4:59 am Post subject: |
|
|
I just noticed this in the server logs from before I rebooted. Seems like it may be related:
Code: | Sep 28 00:26:52 mythback kernel: nfsd: last server has exited, flushing export cache
Sep 28 00:26:52 mythback kernel: klogd 1.5.1, ---------- state change ----------
Sep 28 00:26:52 mythback kernel: Loaded 33765 symbols from 1 module. |
Tom |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 5:23 am Post subject: |
|
|
Interesting. The above was in /var/log/messages, and I see this in /var/log/syslog:
Code: | Sep 28 00:26:52 mythback rpc.mountd[1535]: Caught signal 15, un-registering and exiting.
Sep 28 00:26:52 mythback kernel: nfsd: last server has exited, flushing export cache
Sep 28 00:26:52 mythback kernel: klogd 1.5.1, ---------- state change ----------
Sep 28 00:26:52 mythback kernel: Loaded 33765 symbols from 1 module. |
EDIT: I had the results of showmount here but I was running that from the backend to the frontend, which is obviously backwards. This looks OK:
Code: | showmount -e mythback
Export list for mythback:
/video/tom/v 192.168.1.0/255.255.255.0
/video 192.168.1.0/255.255.255.0
/video/remote_media 192.168.1.0/255.255.255.0 |
Any help would be appreciated...getting desperate here. I'm about out of ideas.
Tom |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 6:11 am Post subject: |
|
|
I'm not sure if this is what I should expect to see from this command (run on the frontend to the backend):
Code: | rpcinfo -p mythback
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100024 1 udp 57306 status
100024 1 tcp 47525 status
100005 1 udp 53615 mountd
100005 1 tcp 34377 mountd
100005 2 udp 46426 mountd
100005 2 tcp 60843 mountd
100005 3 udp 57943 mountd
100005 3 tcp 53509 mountd
100003 3 tcp 2049 nfs
100021 1 udp 43696 nlockmgr
100021 3 udp 43696 nlockmgr
100021 4 udp 43696 nlockmgr
100021 1 tcp 39139 nlockmgr
100021 3 tcp 39139 nlockmgr
100021 4 tcp 39139 nlockmgr |
If anyone has any idea as to whether that look OK let me know.
Tom |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 6:25 am Post subject: |
|
|
I've read about removing entries from /var/lib/nfs/rmtab and tried that...I just get the same error and a line added back into that file. Wow wow wow. Like the subject says. HELP!
Going to look at this again in the morning.
Tom |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 7:40 am Post subject: |
|
|
GOT IT!...Couldn't sleep. A few days ago my world update did this:
Code: | [ebuild U ] net-fs/nfs-utils-2.4.1-r1::gentoo [2.3.3::gentoo] USE="ipv6 libmount nfsidmap nfsv4 tcpd uuid -caps -junction -kerberos -ldap -nfsdcld -nfsv41 (-selinux)" 890 KiB | THAT was the culprit. I discovered that after stumbling on this:
https://www.linuxquestions.org/questions/slackware-arm-108/anyone-having-nfs-trouble-4175658142/
Though that talks about arm. No clue of that. This is on a 32 bit x86 system. I just downgraded the client and server to net-fs/nfs-utils-2.3.3 and all's good. I could about scream. No clue if the issue is client, server, or both. Maybe it's a 32-bit only issue(?). Will worry about logging bugs after I enjoy my weekend.
Tom |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Sat Sep 28, 2019 10:15 am Post subject: |
|
|
https://forums.gentoo.org/viewtopic-t-1101848-highlight-nfs.html
you remind me this, but i couldn't help them in that thread as the issue was glibc version related, and i'm not using this version.
All i could tell you, is that i myself have no issue with it (with nfs3 and 4 shares), and that i suspect nfs-utils to enforce better configuration of the shares rather than a bug.
Code: | [binary R ] net-fs/nfs-utils-2.4.1::gentoo USE="libmount nfsidmap nfsv4 tcpd uuid -caps -ipv6 -junction -kerberos -ldap -nfsdcld -nfsv41 (-selinux)" 0 KiB
Portage 2.3.73 (python 3.6.9-final-0, default/linux/amd64/17.1, gcc-9.2.0, glibc-2.28-r6, 4.14.67 x86_64)
|
|
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6098 Location: Dallas area
|
Posted: Sat Sep 28, 2019 10:24 am Post subject: |
|
|
I don't run nfs at all, but krinn may be right about configuration.
I know that I've had to change configuration items with samba over the years, because of internal changes.
The same with gcc/g++ enforcing things that it was lax about in the past.
I would look at the change log for the newer nfs-utils and see what it has to say vs the one that works. _________________ PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
Back to top |
|
|
mike155 Advocate
Joined: 17 Sep 2010 Posts: 4438 Location: Frankfurt, Germany
|
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sat Sep 28, 2019 1:51 pm Post subject: |
|
|
krinn wrote: | https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e1c9516d14cfe0bf2a9a9b4023780704eed00ecd
Code: | +This fixes the client side error "Stale file handle" when mounting from
+a server running Arch Linux ARM. |
looks more like your issue, coming in -r2 | Thanks!...and thanks everyone for the replies!
Putting that patch in my user patch directory for the existing nfs-utils-2.4.1-r1 on the server side (the client is OK as-is) does in fact fix this. I'm confused as to why everything refers to ARM around this. Both systems in my case are x86. By the way, except for having ipv6 enabled (which wasn't actually intentional), my USE flags are identical to yours.
I see this is related to large file support, which makes sense given that my server is 32 bit, and I do in fact have files in excess of 3 GB in size. I'm sure that's why I ran into this.
Thanks again!
Tom |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Sat Sep 28, 2019 2:36 pm Post subject: |
|
|
tld wrote: | I'm confused as to why everything refers to ARM around this. Both systems in my case are x86. By the way, except for having ipv6 enabled (which wasn't actually intentional), my USE flags are identical to yours. |
I think all systems are affected, must have been found on arm first, but the bug is not trigger on systems where it's default enable |
|
Back to top |
|
|
|