[Solved] User Shares Disappear....only when cache drive assigned


Recommended Posts

My user shares have disappeared.  It is not from an adblocker as discussed in other threads.  The user share do not show up in windows explorer or the webgui.  In another post I thought I discovered that it was tied to a plugin...so I did a clean install of 6.1.6 and it is still happening.  This is stock unraid on the flash with not plugins, dockers or VMs configured (although the image files still reside on the cache drive that I am attempting to assign.

 

Here is happens...if I reboot the server with the cache drive unassigned, all of my user shares are present in the webgui and in windows explorer.  If I stop the array and then assign the cache drive all of the user shares disappear in both the webgui and windows explorer.  If I then unassign the cache drive and restart the array the user shares do not return...however if I reboot with the cache drive unassigned they return...bringing me to the start of this paragraph.

 

Any suggestions.  I have two diagnostics...first grabbed right away from a boot with cache unassigned so user shares present.  Second (1 minute later) is after stopping array and assigning cache then starting array...so no user shares present.

 

Any suggestions?

tower-diagnostics-20160115-1805.zip

tower-diagnostics-20160115-1806.zip

Link to comment

I have the exact same issue since I updated to 6.1.7. While the user shares appears in the CLI they no longer show up in the "User Share" GUI. Like in the original post I do have an extra drive assigned as cache that I don't use for cache purposes. All my drives are XFS formatted including the cache drive.

 

Link to comment

Am I missing something with the adblocker discussion.  I don't believe I have one installed and what makes my situation different is that I am not able to see the files in windows explorer.  I also further noted that I telneted into the console and there are no shares listed in /mnt/user.

 

The problem is still occurring in 6.1.7...just didn't happen on the first reboot after installing it.  I have intentionally rebooted my server a handful of times, there are some instances that don't follow my initial post i.e. sometimes the shares show up like you would expect with all devices including cache assigned.

 

Very confused.  Should I copy my VM and docker image off the cache drive and reformat it?  I have been contemplating moving to BTRFS file system in prep for creating a cache pool.

Link to comment

goinsnoopin,

 

Boot in SAFE mode, unassign cache, and then check its filesystem.

 

I didnt actually boot into safe mode.

But un assigning the cache

then check with

xfs_repair -n /dev/sdd1

then

xfs_repair /dev/sdd1

then reassigned my cache, start array and all my shares returned on the share management page.

 

PS : You can get the device id on the main page after unassigning then check the /dev folder for the numberd partidion

in my case it was sdd1 not sdd.

 

HTH and yes this is very weird

Link to comment

So I backed up my cache drive files and then reformatted my ssd cache drive as btrfs.  Then copied my data back onto the cache drive.  This made no difference.  One thing I noticed was that when I stopped my array then unassigned the cache drive then rebooted, sometimes the cache drive would end up in its slot after the reboot...which I thought was strange.  Shouldn't it remain "unassigned"???

 

So with this in mind I decided to do a new config in case something with the drive assignments was corrupt.  On my first startup with the new config all the user shares were present...I did three test reboots and the user shares did not show up on the subsequent reboots.

 

I also tried something new.  I started the array and user shares were not present...then without rebooting or unassigning cache...I then stopped the array...then restarted...I repeated the start and stop 3 times and by the third time my user shares appeared.

 

Any thoughts?  I thought unraid created the user shares by reading the disks...what would make the third start/stop and different from the two before it?

 

Since I am now on btrfs, I have attached a new diagnostic.

tower-diagnostics-20160117-0829.zip

Link to comment

Just a quick follow up, I am still having this issue with my user shares not showing up. 

 

I decided to try something...I attempted to create a new user share called: NOTWORKING

As soon as I click a button to add the share I get a message in the webgui that says "Share has been deleted" and it never creates.  Attached is my diagnostics....jump right to the end of my syslog and you will see the following...which doesn't look right to me:

 

Jan 20 12:52:14 Tower emhttp: shcmd (88): mkdir '/mnt/user/NOTWORKING' |& logger
Jan 20 12:52:14 Tower emhttp: shcmd (89): chmod 0777 '/mnt/user/NOTWORKING'
Jan 20 12:52:14 Tower emhttp: shcmd (90): chown 'nobody':'users' '/mnt/user/NOTWORKING'
Jan 20 12:52:14 Tower emhttp: shcmd (91): :>/etc/samba/smb-shares.conf
Jan 20 12:52:14 Tower avahi-daemon[2935]: Files changed, reloading.
Jan 20 12:52:14 Tower shfs/user: shfs_readdir: fstatat:  (2) No such file or directory
Jan 20 12:52:14 Tower shfs/user: shfs_readdir: readdir_r: /mnt/disk5/. (2) No such file or directory
Jan 20 12:52:14 Tower emhttp: Restart SMB...
Jan 20 12:52:14 Tower emhttp: shcmd (92): killall -HUP smbd
Jan 20 12:52:14 Tower emhttp: shcmd (93): cp /etc/avahi/services/smb.service- /etc/avahi/services/smb.service
Jan 20 12:52:14 Tower avahi-daemon[2935]: Files changed, reloading.
Jan 20 12:52:14 Tower avahi-daemon[2935]: Service group file /services/smb.service changed, reloading.
Jan 20 12:52:14 Tower emhttp: shcmd (94): pidof rpc.mountd &> /dev/null
Jan 20 12:52:14 Tower emhttp: shcmd (95): /etc/rc.d/rc.atalk status
Jan 20 12:52:15 Tower avahi-daemon[2935]: Service "Tower" (/services/smb.service) successfully established.

tower-diagnostics-20160120-1252.zip

Link to comment

OK...check out the attached screen shot from my putty session. 

 

I go to /mnt/user and type ls and no user shares are listed.  On a whim, I decided to change directories to one of my known user shares "Vdisks" and I get to that path and am able to list the folders.  I then tried a couple other known user shares and I am able to get to these.  Coincidently I am able to go to /mnt/user/NOTWORKING identified in my post above.

 

I also included a screenshot of my Unraid GUI showing that I have no user shares.

 

Any suggestions?  Why don't the user shares show up when I type ls -la?????

 

Dan

unraid_user_share.PNG.95a29a3c7c65cc968982c21daef109fe.PNG

user_shares_gui.PNG.cd82aa6264cbe87de9fe9fd4aace0882.PNG

Link to comment

I did remove all plugins...but did not try the official safe  mode...reason being my usb keyboard does not become active until after the unraid boot menu is gone and I need to dig up a monitor and keyboard to plug directly into the Tower.  I was planning on doing this tonigh.

 

I had another thought...could this be some sort of permissions issue.  Check out the attached screenshot.  Is it normal for the double dot to be root and root.  I thought everything was supposed to be Nobody and Users.  I haven't run the New Perms utility since I converted from unraid 4.7 to version 5 several years ago...maybe I should run it again.

 

unraid_perms.PNG.99d5bb0f756614c52bfaa958605ee69c.PNG

Link to comment

I had another thought...could this be some sort of permissions issue.  Check out the attached screenshot.  Is it normal for the double dot to be root and root.  I thought everything was supposed to be Nobody and Users.  I haven't run the New Perms utility since I converted from unraid 4.7 to version 5 several years ago...maybe I should run it again.

 

Yes. The double dot refers to the parent directory of where you're currently looking. In your example it is /mnt which is owned by uid root, gid root. That is correct.

 

However, your previous example is perplexing. You've effectively done

 

cd /mnt/user

ls -la

 

and found /mnt/user to be empty, though you were then able to

 

cd Vdisks

 

and not get an error. There's something wrong with fuse.shfs here.

 

If you look within /mnt/disk1, /mnt/disk2, etc I take it you can see your user shares as you allocated them but they are clearly not being combined together in /mnt/user

 

What does the output of mount show? Here's mine:

 

root@Lapulapu:~# mount
tmpfs on /var/log type tmpfs (rw,size=128m,mode=0755)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
/dev/sda1 on /boot type vfat (rw,noatime,nodiratime,umask=0,shortname=mixed)
/dev/md1 on /mnt/disk1 type xfs (rw,noatime,nodiratime)
/dev/md2 on /mnt/disk2 type xfs (rw,noatime,nodiratime)
/dev/md3 on /mnt/disk3 type xfs (rw,noatime,nodiratime)
/dev/md4 on /mnt/disk4 type xfs (rw,noatime,nodiratime)
/dev/md5 on /mnt/disk5 type xfs (rw,noatime,nodiratime)
/dev/md6 on /mnt/disk6 type xfs (rw,noatime,nodiratime)
/dev/md7 on /mnt/disk7 type xfs (rw,noatime,nodiratime)
/dev/md8 on /mnt/disk8 type xfs (rw,noatime,nodiratime)
/dev/md9 on /mnt/disk9 type xfs (rw,noatime,nodiratime)
/dev/md10 on /mnt/disk10 type xfs (rw,noatime,nodiratime)
/dev/md11 on /mnt/disk11 type xfs (rw,noatime,nodiratime)
/dev/sdg1 on /mnt/cache type btrfs (rw,noatime,nodiratime)
shfs on /mnt/user0 type fuse.shfs (rw,nosuid,nodev,noatime,allow_other)
shfs on /mnt/user type fuse.shfs (rw,nosuid,nodev,noatime,allow_other)
/mnt/cache/docker.img on /var/lib/docker type btrfs (rw)
root@Lapulapu:~#

 

Note the two lines near the bottom that begin: "shfs ..."

 

Link to comment

Here is the output of my mount.

 

To make thing further perplexing....my VMs on unraid are running (i.e. VMs are in my Vdisks share and they run)  also my plex docker is setup to read my media shares: SageTV and Movies...I am able to play these shows/movies back from plex with no issue. 

 

Dan

 

root@Tower:~# mount
tmpfs on /var/log type tmpfs (rw,size=128m,mode=0755)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
/dev/sda1 on /boot type vfat (rw,noatime,nodiratime,umask=0,shortname=mixed)
/dev/md1 on /mnt/disk1 type xfs (rw,noatime,nodiratime)
/dev/md2 on /mnt/disk2 type xfs (rw,noatime,nodiratime)
/dev/md4 on /mnt/disk4 type xfs (rw,noatime,nodiratime)
/dev/md5 on /mnt/disk5 type xfs (rw,noatime,nodiratime)
/dev/md6 on /mnt/disk6 type xfs (rw,noatime,nodiratime)
/dev/md7 on /mnt/disk7 type xfs (rw,noatime,nodiratime)
/dev/sdf1 on /mnt/cache type btrfs (rw,noatime,nodiratime)
shfs on /mnt/user0 type fuse.shfs (rw,nosuid,nodev,noatime,allow_other)
shfs on /mnt/user type fuse.shfs (rw,nosuid,nodev,noatime,allow_other)
/mnt/cache/docker.img on /var/lib/docker type btrfs (rw)
/boot/config/plugins/dynamix.kvm.manager/domain.img on /etc/libvirt type ext4 (r                                                 w)

Link to comment

Trul,

 

I have tested booting to the offical Safe Mode and can confirm that I have the same situation.  I did discover something interesting while going through this exercise.  I know what I am about to say doesn't make sense...I almost want to make a video to show this.  If I boot the Unraid Tower from a cold start in Safe Mode and start the array, I get the exact same situation as shown in my previous couple of posts.  No user shares in webgui or when I type ls -la in a putty windows, but I can type cd /mnt/user/Vdisks(or whatever the share is) and then ls -la the files and see them.  If I then go to the unraid webgui and stop the array and then click the button for reboot then unraid comes backup and the user shares are present in the webgui and fully visible in the telnet putty window when I ls -la the /mnt/user folder.

 

When I boot to unraid in regular mode...the exact same thing happens...from cold boot no user shares but if I stop array and reboot the user shares show up in webgui and telnet window.

 

I attached a diagnostic for a cold boot while running in safe mode.

 

Dan

 

tower-diagnostics-20160120-1957.zip

Link to comment

Trul,

 

Disk3 is intentionally unassigned.  I pulled the disk from the array a few days ago.  I do plan on replacing it in the near future, but was looking to resolve the missing user share issue first, as I don't want to get into a preclear cycle if someone has a suggestion that requires a reboot.

 

Dan

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.