[Solved] Totally messed up move to Unraid 6! Any advice welcome

[trurl]

Why do you have a user share named "Cache"? Did you create that share on purpose? Sounds like the sort of thing that might have been created by a misconfigured app. Maybe a docker mapping to /mnt/user/Cache when you intended to use /mnt/cache, for example.

[Squid]

36 minutes ago, davekeel said:

Hi Squid - thanks for the instructions - it's started up OK following your guidance and the array has started. I've started to play around with files via my Mac and suddenly all of the folders locked (on a Mac it shows as a red "no entry" sign on the folder) whilst I was looking at them ... and any folder I try to open states "you cannot open 'folder name" because you do not have permission to see its contents". So - more strange behaviour. There are about 2000 folders and they seemed to be slowly locking all of them. Next they are now all appearing as unlocked ... I open one and then close it and they are all locked again?! Totally bizarre! Have attached files as requested! Keeping my fingers crossed that you can make sense of this! Thanks in anticipation

 

tower-diagnostics-20170319-0953.zip

FCPsyslog_tail.txt

I think the bulk of your problem is solely a networking issue.

 

Your network is continually up and down.  Check for loose cables, maybe put it into a different port on the switch.  Reboot switch and router.

 

If / when we get the system stable, you're going to have to reboot it to cancel troubleshooting mode

Edited March 19, 2017 by Squid

[Squid]

11 minutes ago, trurl said:

Why do you have a user share named "Cache"? Did you create that share on purpose? Sounds like the sort of thing that might have been created by a misconfigured app. Maybe a docker mapping to /mnt/user/Cache when you intended to use /mnt/cache, for example.

I think that may have gotten fixed since the screen shot as FCP isn't stating anything about it when the array starts

Mar 19 09:54:31 Tower root: Fix Common Problems Version 2017.03.17
Mar 19 09:54:33 Tower root: Fix Common Problems: Warning: Share Downloads set to cache-only, but files / folders exist on the array
Mar 19 09:54:38 Tower root: Fix Common Problems: Other Warning: CPU possibly will not throttle down frequency at idle

 

[trurl]

37 minutes ago, Squid said:

I think that may have gotten fixed since the screen shot as FCP isn't stating anything about it when the array starts

Mar 19 09:54:31 Tower root: Fix Common Problems Version 2017.03.17
Mar 19 09:54:33 Tower root: Fix Common Problems: Warning: Share Downloads set to cache-only, but files / folders exist on the array
Mar 19 09:54:38 Tower root: Fix Common Problems: Other Warning: CPU possibly will not throttle down frequency at idle

 

Is FCP looking for a user share named "Cache"? The screenshot just said that user share had files on cache when its settings say they should be on the array. Also, diagnostics shows a share anonymized to "C___e" which is cache-only. So maybe he just changed it to cache-only to get the message to go away.

 

My question still stands:

54 minutes ago, trurl said:

Why do you have a user share named "Cache"? Did you create that share on purpose? Sounds like the sort of thing that might have been created by a misconfigured app. Maybe a docker mapping to /mnt/user/Cache when you intended to use /mnt/cache, for example.

 

[Squid]

1 hour ago, trurl said:

Is FCP looking for a user share named "Cache"?

Yup

 

 

BTW, diagnostics includes every .cfg file that's contained within the shares folder on the flash whether or not the share still exists.  So possible that the share did exist, but it doesn't now.

 

And actually, the share name of Cache (OP's error is allowed...  its cache that's disallowed.   Not the greatest thing in the world to do, but it is perfectly valid and will not cause issues as long as disk shares are not enabled (and then you'd run into same share - different case issues with SMB) - FCP does not however check for that edge case issue.

Edited March 19, 2017 by Squid

[davekeel]

Hi Squid- me again! It seemed to work OK for a few hours .. all the "no entry" signs dissapeared and so I started copying a few files/streamed a few files etc etc. Then it suddenly froze... all the files in my mac browser window dissapeared and the server would not respond to shutdown -h now .. had to pull the plug again. Have attached the two files from the flash drive here.... hope you can help! 

tower-diagnostics-20170319-1954.zip

FCPsyslog_tail.txt

[davekeel]

Just seen your responses by the way - thank you. I made the share called Cache (or cache) because I was getting totally confused when trying to first move from 5 to 6 and used some of the 5 config files .. but noticed that whilst I used to have a Cache share, it had gone when moving to version 6. I have since deleted it (I think?!) Sorry I can't be more clear/more technical ... struggling a bit with all of this. Just going to plug flash back in and try a new port for the network and hope that it makes a difference. Meanwhile can you see anything of interest in the files I just posted.. any more clues? Thanks again in anticipation

[Squid]

3 minutes ago, davekeel said:

Hi Squid- me again! It seemed to work OK for a few hours .. all the "no entry" signs dissapeared and so I started copying a few files/streamed a few files etc etc. Then it suddenly froze... all the files in my mac browser window dissapeared and the server would not respond to shutdown -h now .. had to pull the plug again. Have attached the two files from the flash drive here.... hope you can help! 

tower-diagnostics-20170319-1954.zip

FCPsyslog_tail.txt

Still continuing network issues

Mar 19 12:41:01 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:41:01 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:42:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:42:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:43:02 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:43:02 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Mar 19 12:46:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:46:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:46:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:46:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:51:38 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:51:38 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:51:38 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:51:38 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:54:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:54:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 12:54:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 12:54:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Mar 19 13:02:02 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 13:02:02 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 13:02:02 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 13:02:02 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Mar 19 13:03:03 Tower kernel: e1000e: eth0 NIC Link is Down
Mar 19 13:03:03 Tower kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

 

[davekeel]

Thanks - can see what you mean. Have tried a new port and a new cable - wonder if my motherboard is not compatible with 6?! Seem to recall reading somewhere that some MoBo's have ethernet problems?! Just rebooting and will try again and will update in a little while. Really appreciate all the help/advice so far.

 

[Squid]

4 minutes ago, davekeel said:

Thanks - can see what you mean. Have tried a new port and a new cable - wonder if my motherboard is not compatible with 6?! Seem to recall reading somewhere that some MoBo's have ethernet problems?! Just rebooting and will try again and will update in a little while. Really appreciate all the help/advice so far.

 

Its using the e1000 driver, so you've got an Intel NIC so you should have no problems...  I'm still leaning towards cabling / switch issues, but maybe someone else will pipe in.

[trurl]

27 minutes ago, davekeel said:

server would not respond to shutdown -h now

You should use the poweroff command to try to get a clean poweroff if possible.

 

22 minutes ago, davekeel said:

Just seen your responses by the way - thank you. I made the share called Cache (or cache) because I was getting totally confused when trying to first move from 5 to 6 and used some of the 5 config files .. but noticed that whilst I used to have a Cache share, it had gone when moving to version 6. I have since deleted it (I think?!) Sorry I can't be more clear/more technical ... struggling a bit with all of this. Just going to plug flash back in and try a new port for the network and hope that it makes a difference. Meanwhile can you see anything of interest in the files I just posted.. any more clues? Thanks again in anticipation

I was just concerned that something was creating that share.

 

It is very common for people to misconfigure apps in such a way that they accidentally create a top level folder (which as we know is a user share) that is named for a disk. This can be very confusing in Windows since it will make it look like you are seeing a disk when you are really seeing a user share.

 

Or they accidentally create a top level folder (a user share) that is the same as one of their user shares but with the wrong upper/lower case. Since Windows (SMB) makes no distinction between upper/lower case, it will just pick only one of these to show you. This makes people think they have lost files since they are looking at a different user share than they expect, since it appears to have the same name.

[davekeel]

noticed that my mobo has two ethernet ports and so, belt and braces, have connected both of them up to the switch with new cables. Back up and running again but no idea how long for?! Holding my breath!

[davekeel]

11 minutes ago, trurl said:

You should use the poweroff command to try to get a clean poweroff if possible.

 

I was just concerned that something was creating that share.

 

It is very common for people to misconfigure apps in such a way that they accidentally create a top level folder (which as we know is a user share) that is named for a disk. This can be very confusing in Windows since it will make it look like you are seeing a disk when you are really seeing a user share.

 

Or they accidentally create a top level folder (a user share) that is the same as one of their user shares but with the wrong upper/lower case. Since Windows (SMB) makes no distinction between upper/lower case, it will just pick only one of these to show you. This makes people think they have lost files since they are looking at a different user share than they expect, since it appears to have the same name.

 

Thank you trurl - this is really becoming a good education for me - really wish I was a little calmer during the process though. Back up again now with two ethernet cables in play ... will try using it again and keep my fingers crossed.

 

[davekeel]

tentatively putting Solved on this ... my unRaid has now run for almost 20 hours ... performed a full parity check, no ethernet problems reporting in any of the troubleshooting logs, and mangaged to stream a few things without a hic-cup. Still holding my breath - but to everyone who offered such great and helpful advice - thank you so much. Really good to know that such clever people are watching over your shoulder at such times!!!!!