Just upgraded to unraid server pro 4.5.3 all went well. Booted fine. Shares ok. Ran a Parity Check. Completed ok - one drive shows some errors.

 

One of my shares started showing up empty - so I checked the other shares and they were all there.

Then I mapped to the disk and all the data is there as well.

 

So I check the share under the Shares tab and the share was there all configured correctly. Clicked on apply for an extra measure of confirmation. Guess what? ALL MY SHARES are now gone.

 

Then I tried to create the shares from scratch does not seem to be taking. I can not seem to create a share.

 

Could someone also point me to a discussion on why I should or should not use NFS shares - I have not been using them thus far.

 

Thanks.

Ok guys I did a server reboot - and the shares are back.

This is the first time I have had any kind of strange behavior with my unraid server. The only change is the upgrade to 4.5.3.

Personally I'm staying with v4.5.1 as it is fewer less complaints compared to v4.5.3. I had problems with v4.5.3 was that the S3 script placining my server to sleep, suspending disks after the set time their suppose to suspend would work and not work and WOL stopped working too, all because of v4.5.3. I'd downgraded back to v4.5.1 as it seems (well for me) a better release with no issues on my server now. 

All the shares except for the flash drive have not disappeared again. BUT I can access everything via the \disk1 \disk2 etc shares.

 

Ok I guess I got to down grade to 4.5.1 wonder why everyone is all quiet about this problem. No one else is seeing this?

If you stop/start the arrays via emhttp, they should show up if they were missing before. The reason I haven't noticed it is my server has long uptime so once it's working it remains working. The reason I haven't complained is a simple toggle of the array 'fixes' the issue.

Yes a stop and start of the array seems to fix it.

This will be a pain.

 

BRIT which version of the server are you running?

I'm using the md device driver and emhttp from 4.5.3 installed on top of a Slackware-Current (13.1) distro. Not exactly a typical setup.

Are you still running with 4.5.3 with the disks going missing? Have you downgraded to 4.5.1 yet? I would for sure. It would seem to be a pain in the a.. to manually stop/start the array just to get this problem fixed.

 

Yes a stop and start of the array seems to fix it.

This will be a pain.

 

BRIT which version of the server are you running?

It's not a pain in the ass at all. It's a 4 second fix. Browse to the main page of emHTTP, click stop array ... wait ... click start array. That only has to be done once per bootup. In my case that means once a month.

Yeah point taken BRiT, though not everyone reboots their server once a month or leaves it on 24/7. In your case, yeah its not a problem to simply stop/start the array, but for those who power off and on their servers daily basis (like myself) or even multiple times a day, you could see it would be a inconvenience to do this as it just should work.

It's not a pain in the ass at all. It's a 4 second fix. Browse to the main page of emHTTP, click stop array ... wait ... click start array. That only has to be done once per bootup. In my case that means once a month.

 

Ahhh if I only have to do the stop start array once after each reboot - I agree not a big deal at all! I never shut down my server so only reboots are if there is long lived power outage or an OS upgrade.

 

I have not downgraded yet - problem has not come back - so I guess the one stop start seems to have worked.

 

Any idea why this happens?

I had a similar issue in 4.5 to 4.5.1. I ended up adding a "sleep 1m" in my go script so the array fully started before any of my addons started (which was causing my user shares to not mount correctly).

Could you be more specific exactly where you put the "sleep" command? In rc.local before emhttp? After emhttp?

 

Thanks,

 

Phil/TW

I put it after emhttp but before unmenu and any of my other addons.

Ok this is getting annoying - so its not just a stop restart once after a reboot - I have now done a stop restart like 3 times - there is obviously something bad going on. My server stays up all the time. Realizing a user share is down and then having to stop and restart is a pain.

 

What's the last known good version of Unraid? the previous version I ran had the windows copy problem "too many files open" that's the only reason I upgraded and now I have this problem.

 

Any suggestions on which version to downgrade to? 

What should I do to down grade to 4.5.1? Which files need to be replaced?

 

Thanks in advance.

What should I do to down grade to 4.5.1? Which files need to be replaced?

 

Thanks in advance.

only two

 

bzroot and bzimage

I down graded from 4.5.3 to 4.5.1

 

But shares are still disappearing.

 

The primary shares like \\disk1 or \\disk3 are fine

but the user shares like \\dvd or \\music keep disappearing. Its VERY ANNOYING!

 

HELP

Anyone? Any help?

 

Is no one else having this problem with 4.5.1?

When do they disappear? Is it on reboots? Is it resuming from S3? Is it after sitting idle? Is it after doing something on the server? How are you detecting they disappear?

I normally have my server running all the time - ie: I only reboot for upgrades or on a drive replacement.

 

Since I upgraded to 4.5.3 the shares would disappear every week or so. I down graded to 4.5.1 everything was fine for about a week, and then when I map a drive to a share the user share shows empty - ie: no files. However if I map to drive1\usershare everything is fine. Stop and start in the console and all the shares come back up. I last stopped and started the server on 5/22/10 so far shares are there.

 

I never had any of these problems till I upgraded to 4.5.3 - the only reason I did the upgrade was because I was getting the too many files open error with windows 7.

Just to clarify the user shares come up on reboot.

Work for awhile.

Then disappear after awhile.

 

To be precise the shares are there - but the files do not show.

However when you map via disk1\usershare everything is there.

Stop start in console and all comes back.

I know that this might not be the correct direction but have you tryed 4.5.4, i beleave that tom did do some more fixes then just the format issue. it might be related.... I also was geting a simalar issue when I built a System with a Nvidia 8200 chipset. It did not like ahci kind of wierd. The system had other issues so i rebuilt it with another motherboard with a intel chipset and all problems went away.

Since I upgraded to 4.5.3 the shares would disappear every week or so. I down graded to 4.5.1 everything was fine for about a week, and then when I map a drive to a share the user share shows empty - ie: no files. However if I map to drive1\usershare everything is fine. Stop and start in the console and all the shares come back up. I last stopped and started the server on 5/22/10 so far shares are there.

 

Daytona, the next time the user shares disappear, can you perform the following debugging steps?

• df
  
• ps -ef
  
• free
  
• cp /var/log/syslog /boot/config/sys.log
  
• cp /var/log/messages /boot/config/msgs.log
  

 

The logic behind these steps are to see if 'shfs' is running and shows up under /mnt/user as well as to examine what processes are running, what error/warning/info messages are in the log files, and how your memory is used.

 

I'm going to take a wild stab in the dark here and guess that something is killing the 'shfs' process which provides the unified user share view.

Since I upgraded to 4.5.3 the shares would disappear every week or so. I down graded to 4.5.1 everything was fine for about a week, and then when I map a drive to a share the user share shows empty - ie: no files. However if I map to drive1\usershare everything is fine. Stop and start in the console and all the shares come back up. I last stopped and started the server on 5/22/10 so far shares are there.

 

Daytona, the next time the user shares disappear, can you perform the following debugging steps?

• df
  
• ps -ef
  
• free
  
• cp /var/log/syslog /boot/config/sys.log
  
• cp /var/log/messages /boot/config/msgs.log
  

 

The logic behind these steps are to see if 'shfs' is running and shows up under /mnt/user as well as to examine what processes are running, what error/warning/info messages are in the log files, and how your memory is used.

 

I'm going to take a wild stab in the dark here and guess that something is killing the 'shfs' process which provides the unified user share view.

One more to also run when the shares dissapear

smbclient -N -L localhost

 

It will let you know which shares are being advertised by your server.

 

Joe L.