Unraid Shares Disappear

[jacobgad]

Hi everyone,

 

I have been having a problem where about once a week my user shares disappear from unraid's share tab and become inaccessible from the network. The resolution each time is to stop the array and then restart it or just reboot the server but this seems to be occurring approx every week.

 

the annoying thing is that i need to download new metadata for plex every time this happens and on a large library this becomes really annoying really fast.

 

I have attached the diagnostics zip,

Any help at all would be greatly appreciated

thank you  

 

ungad-diagnostics-20171205-1122.zip

[jonp]

We need a copy of the diagnostics from the system AFTER it has gone into the state where the shares disappear.  If you can attach that, it would be very helpful in diagnosing your issue.

 

All the best,

 

Jon

[Squid]

I actually thought it was after due to all the "Too Many Files Open".  But, since the syslog has been rotated too many times and is now truncated a reboot, wait for the problem to reoccur then immediately posting a new set of diagnostics would help because we'd be able to see what happens prior to all those errors occuring

[jacobgad]

Thank you so much for your help, Jonp and Squid. it may be a week or so before the issue happens again. but as soon as it does i will get a copy of the diagnostics when it is happening and after a reboot of the array.

thank you again, 

[jacobgad]

I have a copy of the diagnostics from the system AFTER it has gone into the state where the shares disappear,
thank you in advance for your help 

 

ungad-diagnostics-20171217-2300.zip

[Squid]

I think this is what is happening.

 

You've got mover running every hour.

 

Because of the logging done by mover, it has very rapidly filled up /var/log.  Under settings, Scheduler, mover settings, disable mover logging.  If nothing else, it makes for a nicer syslog to read.

 

Mover keeps attempting to move appdata from the array to the cache drive.  All well and good, but it is continually unable to do so completely because you've got docker running.  Settings, docker, stop running.  Then Main, Array Operation, run mover.  After its done, re-enable docker.  Now everything should be on the cache drive and mover won't complain as much, and as a bonus you'll see a big speed bump.

 

Of course, you're going to have to reboot first to accomplish this as with /var/log filled, strange things result.

[Squid]

Additionally, your cache is currently at 85% full.  You might want to consider setting your shares (except for appdata and downloads shares) to bypass the cache drive altogether.  You won't see a huge decrease in access time for media files, but you will give yourself some more breathing room which again will make the system more responsive altogether as appdata and downloads will tend to stay on the cache drive instead of falling back to the array.

[jacobgad]

Thank you for the reply Squid,

I stopped docker and and let the mover run successfully as well as well as reduced the number of timed the mover will run.

Unfortunately the problem is still going on, any further help would be greatly appreciated.

 

i have attached a new diagnostic file from when the server's shares have disappeared.

ungad-diagnostics-20171223-0540.zip

[Squid]

On 12/17/2017 at 8:49 AM, Squid said:

you're going to have to reboot first to accomplish this as with /var/log filled,

You didn't do this step.

[jacobgad]

Hey squid, i thank you again for all your help. sorry for the late reply i was on holidays. 
I don't understand what step i have missed. Sorry to ask but could you explain it again as I don't follow.

 

so far i have disabled docker and vm's then run the mover, wait for that to complete then enable docker and vm's again.

thank you again for your help, and i apologise if i am not understanding correctly. 

 

[JonathanM]

23 minutes ago, jacobgad said:

I don't understand what step i have missed.

 

On 12/17/2017 at 8:49 AM, Squid said:

reboot first

 

[caballopazo]

having the same problem now. i got home and my shares are no longer showing on unraid.

my dockers and VM have disappear, if i try to reboot or power down, it doesnt work, i have to hard reboot.

below is my diagnostics, also a screen shot of the log error i'm getting

 

tower-diagnostics-20180824-0730.zip

 

[Squid]

Your cache drive dropped offline.  Most likely a poor connection on the sata/power connectors.  Reseat.

 

Additionally, disable mover logging (settings, scheduler, mover settings) - serves no real point, but just takes up log space.

[caballopazo]

thank you Squid.

i was able to check all the wires, and harness to power supply was off.

all came back ok, however, i now have a brand new disk i install two weeks ago showing me a red x. smart test is showing in good shape.

i'm not sure why is not given me a green light.

here's the log and diagnostics, and also a screen shot of the system log

 

tower-diagnostics-20180824-1934.zip

 

[caballopazo]

hi everyone i followed these steps thank you all for the support. i will check back once the parity finishes rebuilding.

 

 

[Squid]

50 minutes ago, caballopazo said:

all came back ok, however, i now have a brand new disk i install two weeks ago showing me a red x. smart test is showing in good shape.

 

Disks get disabled only if a write to it failed.

51 minutes ago, caballopazo said:

i was able to check all the wires, and harness to power supply was off.

I would guess that the odds are decent that in the process of fixing the power to the cache drive, you slightly knocked out of place the sata / power connector to your other drive (why I'm a huge fan of hot-swap cases - even the cheap ones are worth their weight in gold).

[caballopazo]

agreed .. will need to think to update!!!

have a great weekend and thanks for your help!

[muslimsteel]

I just had this issue, took a diagnostics file right after it happened. I was working on some files in krusader when suddenly it said the directory did not exist anymore. I went up a level and that one did not exist either. Then I looked at my shares and they were gone. Googled it and came across this post. So I took some diagnostics and then rebooted. Server seems to be ok now, but just seems odd, was wondering if there was anything in the diagnostics that would suggest the problem. Thanks!

hulk-diagnostics-20200610-2159.zip

[mbc0]

Hi,

 

Did this ever get resolved? I have had the same problem 3 times now, each time if I browse to /mnt/user/ it is empty! reboot and all is well again, the has happened 3 times in the last 36 hours, I have attached my diagnostics if anyone can help please?

unraid1-diagnostics-20200819-1739.zip

[Squid]

2 hours ago, mbc0 said:

Hi,

 

Did this ever get resolved? I have had the same problem 3 times now, each time if I browse to /mnt/user/ it is empty! reboot and all is well again, the has happened 3 times in the last 36 hours, I have attached my diagnostics if anyone can help please?

unraid1-diagnostics-20200819-1739.zip 221.73 kB · 1 download

Are those diagnostics from a state when /mnt/user is empty?

[mbc0]

16 minutes ago, Squid said:

Are those diagnostics from a state when /mnt/user is empty?

yes they are 🙂

[trurl]

1 hour ago, mbc0 said:

yes they are 🙂

Those diagnostics are saying you have 63TB of files in user shares, with lots of user shares with files on the disks.

 

Possibly you are having a browser issue instead. Do you have an adblocker or anything else that might interfere with accessing the webUI? If so, whitelist your server.

[mbc0]

11 hours ago, trurl said:

Those diagnostics are saying you have 63TB of files in user shares, with lots of user shares with files on the disks.

 

Possibly you are having a browser issue instead. Do you have an adblocker or anything else that might interfere with accessing the webUI? If so, whitelist your server.

Nothing to do with browsers, I logged onto the unraid server via putty and direct and all shares were gone.  At the moment it is looking like the tdarr docker was responsible as I was able to reproduce the problem by rescanning files.   The docker owner has made some changes last night and am re-running, so far so good!

[mbc0]

Just happened again.. Still could be the tdarr docker causing this?  I once again have just connected directly onto the server and all shares gone when viewing through putty, diags attached again just after it happened. Stopping & Starting the array has cured it.

unraid1-diagnostics-20200820-1630.zip

[trurl]

Those diagnostics are still showing you have user shares. Did you take them before stopping/starting? Seems unlikely that a container would cause this.

 

Just to make sure there isn't some misunderstanding, what are you looking at from putty that shows your user shares are gone?