I'm powering the unit on/off as directed and using a APC SMART UPS.  Everytime the power is cycled for reboot or power off to add a drive all settings are lost except the array config with shares.  unRAID thinks it needs a parity check and reads out 15677 days since last parity check.  Parity is good and all drives are aligned.  I'm runnung b12a and everything is green.I want to upgrade but not until I figure out what I've done to cause this behavior.  The syslog resets each reboot too.  The last time was this am my time.

 

Anyone seen or heard of this type of behavior? 

 

Syslog attached

 

Thanks

syslog_seanant_May302012.zip

What type of install/hardware?  i.e. is it a virtualised or physical server?  Do you have any plugins and do they persist through reboot?  Can you attach a syslog after a fresh boot?

The attached syslog is what is logged every time I reboot.  The last time was yesterday am.  I'm running a supermicro x9scm-f with 2 1015s and 8 drives.  Plugins running - Plex and Headphones.  I was using unmenu until the last update caused the array to be in "starting" state and never completely come up.  I deleted unmenu and now running with this problem.

It looks like some of your unmenu plugins are still sticking around - for example, python 2.6 (installed by your couchpotato addon) is being removed in favour of an earlier version (2.4) - being that it is located in /packages, it looks like it was selected from unmenu previously.

 

I can't see anything else obviously wrong (perhaps greater minds than mine will!), but I'd be tempted to back up your /config/go file, then remove "/boot/unmenu/uu" and "cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c " if they're still in there.

exactly HOW are you powering down your server?

 

Please list the steps... in detail.

Thanks. I power down from the web GUI and select shutdown and click ok with yes checked. JFK use the cli I use halt.

 

Both options seem to shutdown gracefully

Thanks. I power down from the web GUI and select shutdown and click ok with yes checked. JFK use the cli I use halt.

 

Both options seem to shutdown gracefully

From the command line, "halt" is NOT going to stop the array cleanly and will force a parity check upon reboot.

 

As delivered from lime-tech, the only possible clean shutdown command is "powerdown"  (and even it depends on the unRAID management interface working properly)

 

from the unRAID management interface you must first press the "Stop" button to cleanly stop the array, then you can power down.

 

I think the reason for your settings not being saved is you are not cleanly stopping the array first.

 

Joe L.

I tried this and get the same result.  I forgot to mention I run a stop script which should work?  I see the drives unmount and to me looks like things are gracefully shutting down.  The process takes a monute or two.  There's no abrupt off.  Will this cause parity check?  If I'm going through the wrong process I will change what I do.  If there's another underlying issue I want to figure it out and get back to online status. :-)

Shutdown only using the the unRAID GUI. The "powerdown" add-on allows clean CLI shutdown.

heve you even looked in your syslog?

 

May 30 06:51:07 Tower apcupsd[9797]: /sbin/apcupsd: Bogus configuration value (*invalid-cable*)

May 30 06:51:07 Tower apcupsd[9797]: Terminating due to configuration file errors.

May 30 06:51:07 Tower apcupsd[9797]: Lock file read error. ERR=Is a directory

May 30 06:51:07 Tower apcupsd[9797]: apcupsd error shutdown completed

 

May 30 06:51:13 Tower useradd[11946]: failed adding user 'istat', data deleted

May 30 06:51:13 Tower groupadd[11947]: group added to /etc/group: name=istat, GID=1000

May 30 06:51:13 Tower groupadd[11947]: group added to /etc/gshadow: name=istat

May 30 06:51:13 Tower groupadd[11947]: new group: name=istat, GID=1000

Your share names "Printer Driver" and "Window Images" are causing all kinds of issues as the space in them causes commands to fail:

May 30 19:00:16 Tower emhttp: shcmd (211): echo '/mnt/user/Printer Driver -async,no_subtree_check,fsid=100 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports

May 30 19:00:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/Windows Images.cfg: No such file or directory - assigning defaults

same with other share names with embedded spaces.  (replace the space with an underscore or a dash and it will be much better)

May 30 19:32:06 Tower emhttp: get_config_idx: fopen /boot/config/shares/Juniper Customers.cfg: No such file or directory - assigning defaults

May 30 19:32:06 Tower emhttp: get_config_idx: fopen /boot/config/shares/Linux Images.cfg: No such file or directory - assigning defaults

 

These have nothing to do with it needing a parity check upon reboot, but I strongly suspect you are not stopping the array AND WAITING FOR IT TO STOP before powering down.

Thanks.  Yes I've read through the syslog.  My problem is I don't know where to start when researching an error in the syslog.  I've got so much data I'm cautious and fear what could happen if I play around with the cli too much.  I don't have a test unraid server.  I've had numerous hardware failures the last 12 months due to what sounds like neglect on my part. The power outage last october hit me hard and most all of the server has been replaced with new equipment since then.  Ive got 3 origional drives from when I first started this server everything else is new.  I ask now before I delete or modify if I'm not 100% positive.  Thanks for taking time to read through my log.  I do appreciate the help.

R/-S

Thanks.  Yes I've read through the syslog.  My problem is I don't know where to start when researching an error in the syslog.  I've got so much data I'm cautious and fear what could happen if I play around with the cli too much.  I don't have a test unraid server.  I've had numerous hardware failures the last 12 months due to what sounds like neglect on my part. The power outage last october hit me hard and most all of the server has been replaced with new equipment since then.  Ive got 3 origional drives from when I first started this server everything else is new.  I ask now before I delete or modify if I'm not 100% positive.  Thanks for taking time to read through my log.  I do appreciate the help.

R/-S

Are you stopping all your add-ons before attempting to stop and power down your server?  Disks cannot be un-mounted if they are busy, and they are busy if processes are using them in any way.

Those proecesses must be stopped first, then the array will write to the flash drive the fact it has stopped cleanly.

 

As far as the user-share names... just rename the...  Juniper Customers would become "Juniper_Customers", etc...

 

 

Aren't those first few errors just a product of removing unmenu incorrectly, as he stated earlier?  In which case, stopping them from starting by removing the line from the go script as I mentioned above should resolve that. 

 

That helped. 

 

 

Helped as in fixed or..?

it cleaned up the mess a little.  It looks like the settings are being wiped, the syslog is new.  I havent tried anything else. 

The syslog still shows add-ons being installed. Use the factory original stock go file (see attached). Backup the current go file or just rename it "/boot/go.old" and use the attached file. The attached should be renamed as "/boot/config/go".

go.txt