I have been having endless trouble with my server crashing after being on for about ten minutes. In desperation i decided to go from scratch, i didnt take a screen shot of my drives as the disks all have stickers on the with the disk numbers and sdb, sde etc etc, thats how they were..now the prefix of sdb,sdc etc have altered from the disk numbers..so for example where i had disk 2 labelled as sdg it now prefixed in the assignment page as sdd..so do i assign it as per the disk number or go by the prefix ?..i initially went by disk number as per the disk labels so this means the prefix is all wrong, like its shifted..it says configuration valid and started a parity sync which took 8 hours, when i came to check it it was all locked up still and had to hard reset it loads of times..it wouldnt boot so i left it untill this morning and its booted up straight away but says that parity is invalid...does this mean it will have written over my parity drive now so no point trying to write the back up of the flash back to flash device to see if it boots ?? not sure how i should proceed as i dont want to loose my data.

Not quite sure what state your array is in at the moment, so not sure how to tell you to proceed.

 

You should never rely on the sd? designations as they are generated dynamically by Linux at boot time depending on the order that disks are recognised.  unRAID from v5 or later recognises disks by their serial number - not their sd? designations. 

 

Having said that it is rare for the sd? designations to change across boots so if that is happening to you on a regular basis there is some other factor at play.  For instance are you sure that your power supply is sufficient for the number of drives you are trying to use.

Its a 700w running 6 drives and never had any problems with it before, the sd? designation has not changed before, its the first time its happened but like i said, i assigned it as per the disk number i have them labelled as not by the sd? on the labell. not sure if i have corrupted my parity drive now..i left it syncing last night and got locked out (this is the trouble i have been having that has caused me to do this, server keeps stagnating on me so i have to hard boot it to get in), so i went from scratch...but the sd? assignments are now wrong..i assigned the disks by thier serial numbers and disk numbers, ie 1,2,3,4...if these are not right then the parity will show as invalid after a sync ?

 

EDIT: I have just now reformatted my flash drive and done a fresh install, its booted up and seems to be a lot nippier, bizzarly the drives awaiting to be assigned are in the correct slots already, im adamant that i have assigned the drives correctly as i went by the disk number and serials, parity is still showing as invalid..will this correct itself ?

 

Is it all possible that the sd? assigment changed has been caused by the constant hard booting i have to do ??

It does sound as if your power supply should be sufficient.

 

It is always worrying when you are getting lockups.  It is normally hardware related in some way but can be hard to track down.  It is probably worth selecting the memtest option from the boot menu and running a test overnight.  RAM issues can produced all sorts of unpredictable errors.

 

Parity will always be invalid on a new setup until you successfully complete a parity sync.

 

A syslog might give some indication as to why you are getting server crashes.  Since a crash precludes you getting a syslog it might be worth leaving a log window open the next time you try a parity sync to see if anything useful shows up at the end.

Im leaning toward hardware issue... pretty sure psu is good, i can leave it doing a mem check i suppose but i have 16gb in dual channel mode (4x4gb)..might be worth just dropping two dimms out and see how the other two fare. I have ONE OF THESE brand new ready to test with but the copy of my licence is on the array..how can i go about assigning it to this new key as it will not boot with all my drives in unless the serial is correct for my licence

i just read your post on my other thread...that is going to be a royal pita...i dont think its the flash drive then, i have booted server fine with other flash drive and this one is booting no problem..i just have to keep hard resetting because of the lock ups...because i have done a fresh install after a hard reboot the drives are all coming back unnassigned...and i have to re allocate them..im going to leave it in this state (unassigned) for a hour or so and see if it still locks me out as it seems to be quite fluid when the drives are not assigned.

i just read your post on my other thread...that is going to be a royal pita...i dont think its the flash drive then, i have booted server fine with other flash drive and this one is booting no problem..i just have to keep hard resetting because of the lock ups...because i have done a fresh install after a hard reboot the drives are all coming back unnassigned...and i have to re allocate them..im going to leave it in this state (unassigned) for a hour or so and see if it still locks me out as it seems to be quite fluid when the drives are not assigned.

If it locks up with no drives then I think you must have a hardware issue.

 

I noticed you have 4 RAM sticks installed.  Many motherboards are more stable with only 2 installed so it could be worth trying that.

 

Another thing I can think of to check is that the fans are all working fine and you are not getting over-heating issues.  This would apply in particular to the CPU fan.

Ok..i left the server on without the array started..no lockup at all and thats been three hours now..i am now going to assign the drives and start array..parity sync will start automatically..watch this space...

 

EDIT: A few screen shots if they help, seems to look ok other than the faulty radio button ?? is that because its not written parity yet ?

 

EDIT:2 So two hours into parity sync with six hours to go and it remains fluid thus far.

 

 

 

You will always get a parity check if you don't stop the array before rebooting or powering down, like you have been doing. That is by design since parity might not have been correctly updated.

 

Also, you mention having a backup of your flash drive. That backup should always be made with the array stopped, since booting from a backup that had the array running will make unRAID think it wasn't stopped the last time.

 

And don't make the mistake of using a flash backup that does not have your current disk configuration on it. Someone did this after they had reused their parity drive as a data drive, and unRAID started writing parity to it.

 

Any time I make any change to the array like replacing or adding a disk I always stop the array and copy the flash share to my PC then restart the array.

 

You can often safely shutdown from the command line if the webGUI is not working. See the powerdown plugin thread.

Thank you for those tips..quite possibly i have backed up my flash whilst array is running so something to bare in mind. i dont think its relevant in this case as i worked out which drive is which as i labelled them.

 

The worry i have now is that the parity check has slowly moved to 81 days and has only moved from 32.8% to 32.9%

 

im going to buy a 3 or 4 tb drive tonight to arrive for tomorrow awaiting any verdicts...im thinking the parity drive is faulty ??

 

EDIT:.. just noticed the hanging seems to be when its syncing parity from disk1, i just clicked it to try and see whats on it and its crashed on me..locked out again...if its the disk 1 that is screwed (all disks were reading healthy) then im screwed as i have started to write parity from a damaged disk ??

Thank you for those tips..quite possibly i have backed up my flash whilst array is running so something to bare in mind. i dont think its relevant in this case as i worked out which drive is which as i labelled them.

 

The worry i have now is that the parity check has slowly moved to 81 days and has only moved from 32.8% to 32.9%

 

im going to buy a 3 or 4 tb drive tonight to arrive for tomorrow awaiting any verdicts...im thinking the parity drive is faulty ??

 

EDIT:.. just noticed the hanging seems to be when its syncing parity from disk1, i just clicked it to try and see whats on it and its crashed on me..locked out again...if its the disk 1 that is screwed (all disks were reading healthy) then im screwed as i have started to write parity from a damaged disk ??

Whenever parity is being created/checked then unRAID will be trying to read from all disks simultaneously.  It is possible that unRAID is having problems reading from one or more disks and this is slowing down the parity sync.  Posting a syslog covering a period where the sync is going slowly may allow others to comment on whether any disc issues are being logged.

... just noticed the hanging seems to be when its syncing parity from disk1...

This is not the way parity works.

 

How does parity work? If you understand this it will go a long way in preventing mistakes when making changes to your array.

maybe i am misinterpreting 'parity sync', given the fact this is a fresh virgin install and i am 99% i have assigned the drives correctly going by the the labels i applied to to each to identify visually which one is 1,2,3 etc, ignoring the sd? prefixes which no longer tally up with the sd? i wrote on the labels next to each disk number, am i right to say that it is not writing to the parity disk, merely just trying to verify it ?..no disks have been physically removed,replaced or plugged into a different controller through any of this.

The parity check triggered by an unsafe shutdown is a correcting parity check, which means it will rewrite any parity that is incorrect.

Well..somethings afoot. i have just had to hard reset it, of course it wants to do a parity sync..i stopped it for now, i can access all the files on each disk so i dont understand whats causing the server to keep falling over 

Run memtest man and give it 24hrs, sounds like a hardware issue. Best to rule out hardware first and narrow down the issue, also next time screenshot your server, never work from mount points, work by the drive serial numbers and record what drive serial number has what data on it, drive sn: 108987345 parity for example. If you move sata cables around or the controller resign them or you disable a controller in bios then your lost.

thats essentially what i have with the labels on the disks, almost like a hard copy as each is labelled by its disk number and serial..dont understand why the sd? designation has shifted along by one..i never moved any drives, a couple got unplugged so i could read the serials but the array was in a stopped state and drives plugged back to same ports, its impossible to plug the drives into different cables as the wiring is so tidy, it wont physically plug into another drive.

so i have just turned on server this morning..to try and get this parity sync done..im getting very concerned now as the array is unprotected. is the log at the top of the page the log i would need to post or is there a different one hiding on the flash i need to get hold of for someone to see.

 

EDIT: cant even do that...after a couple of minutes navigating the folders on the disk via unraid its locked up..so i can emulate the problem by doing this but i cant grab a log from it once its done it as it locks me out, removed the flash drive to look on there but cant see it..i dont think its going to show anything as it does nothing once it locks out.

 

i have four dimms in this machine, i am going to pull out two and then try the other in different dimm combinations, this should be a short cut to a memtest for the time being

This is going from bad to worse, i moved the dimms around and eventually put them all back and it wont boot at all from the usb stick after i tried that in a different port

This is going from bad to worse, i moved the dimms around and eventually put them all back and it wont boot at all from the usb stick after i tried that in a different port

Did you heck what device the BIOS is set to boot from?  There is a good chance that the BIOS decided to be 'helpful' and reset the boot device order.

 

I am assuming that the system is competing its 'post' testing so is seeing the DIMM's OK.    I would definitely try with only 2 installed as that is likely to be more stable if there are any timing related issues on the RAM.

Definitely do a memtest before continuing. Memtest is available for selection on the boot menu when you first boot unRAID.

 

If that comes out OK then there are several ways to get a syslog here. Use the "tail" method.

Ok..sound advice. I had a look at the sys file on the flash drive...If I make a backup know up of it can I just move the mem test entry to be the default boot option ?

Ok..sound advice. I had a look at the sys file on the flash drive...If I make a backup know up of it can I just move the mem test entry to be the default boot option ?

yes

Well.... what can I say. Coincidence or cause of the trouble ?. As my my unraid usb died I temporarily used a different one which Tom kindly transferred my key to and now it firing on all cylinders... The issues of the parity drive saying it needs to sync have disappeared and it has completed a parity check without errors... server has not crashed once either !!!... guess this is solved.