View Full Version : Mounts are gone after this mornings update

5th October 2017, 02:26 PM
I have F26 on several boxes and do daily updates. After this mornings update & restart I no longer have any mounts that worked perfectly as recently as 2 minutes before the update & restart. The mounts were created by entries in /etc/fstab with the format //ip/share /mnt/dir cifs credentials=/home/user.smbcredentials,uid=theUser,gid=theUser.

I assume something in the OS update made things more secure. We are using qnap NAS for storage and haven't had any issues in years. Any ideas on how I need to mount common shares in F26 as of this morning?

Update: There were no changes on anything else on the network. Everything is local to each other, no firewalls or other impediments. I can connect to the shares via GUI browsing to the NAS and share (SMB), entering the same credentials I used in the credentials file.

5th October 2017, 11:47 PM
I have the same issue. If something is no longer supporting cifs we should have known ahead of time not a surprise.
I found if on boot you tell fedora to boot the previous version your mounts come back. I am halting all updates until this issue gets fixed.
Anybody know what to do to fix this?

I can still access the shares in nautilus under other locations and network manually

6th October 2017, 06:23 AM
you may want to revert the kernel update and add a line to /etc/dnf/dnf.conf to prevent updates until this is fixed.

cat /etc/dnf/dnf.conf


6th October 2017, 09:45 AM
I too have lost the auto mount for my NAS under both Fedora 27 Beta with kernel 4.13.4 and Fedora 26 with kernel 4.13.4. If I boot either to a previous kernel the auto mount works.

Seen in journalctl for last boot (Fedora 27 Beta)

Oct 06 00:46:45 lxde27 kernel: FS-Cache: Loaded
Oct 06 00:46:45 lxde27 kernel: Key type dns_resolver registered
Oct 06 00:46:45 lxde27 kernel: FS-Cache: Netfs 'cifs' registered for caching
Oct 06 00:46:45 lxde27 kernel: Key type cifs.spnego registered
Oct 06 00:46:45 lxde27 kernel: Key type cifs.idmap registered
Oct 06 00:46:45 lxde27 kernel: No dialect specified on mount. Default has changed to a more secure dialect, SMB3 (vers=3.0), from CIFS (SMB1). To use the less secure SMB1 dialect to access o
Oct 06 00:46:45 lxde27 kernel: CIFS VFS: cifs_mount failed w/return code = -112
Oct 06 00:46:45 lxde27 systemd[1]: mnt-buffalo.mount: Mount process exited, code=exited status=32
Oct 06 00:46:45 lxde27 systemd[1]: Failed to mount /mnt/buffalo.
Oct 06 00:46:45 lxde27 systemd[1]: Dependency failed for Remote File Systems.
Oct 06 00:46:45 lxde27 systemd[1]: remote-fs.target: Job remote-fs.target/start failed with result 'dependency'.
Oct 06 00:46:45 lxde27 systemd[1]: mnt-buffalo.mount: Unit entered failed state.

If I add vers=1.0 to the mount line in /etc/fstab I can manually mount the NAS. For example:

// /mnt/buffalo cifs vers=1.0,username=xxxxxxxx,password=yyyyyyyyyyy 0 0

6th October 2017, 10:19 AM
Thought I remembered seeing something about this:

6th October 2017, 10:25 AM
OK. Version 1.0 = bad. I don't need to mount the NAS from anywhere but my CentOS 7 server, and that currently works fine.

A fix will be greeted with open arms so that I can mount the NAS from elsewhere should I feel the need.

12th October 2017, 07:48 AM
I have the same problem here, and yesterday's kernel update seems to have made things worse. Until yesterday, I wasn't able to mount my cifs shares on my nas-server and my internet router, but I could mount a cifs share on my raspberry pi. Since yesterday, I cannot mount any cifs share.