having successfully merged my previous 2 snapraid machines into one new one with dual parity. a sync-H plus a check all reported no errors. On powering up the machine this morning one of the data disks failed to mount and had a corrupted partition table, I have rebuilt the table using gparted and started a Fix -l for just the dead disk (d3).
I then aborted this as I had not appreciated how much space would be required for the log file and dont think it will fit in the space available. It appears to be using approx 100mb for each 1% and with 3TB to be recovered will be very tight on the 3gb free space on my boot drive, which is only a 16gb USB3 sandisk Extreme (to save on sata ports)
My options/questions are
1) can I change the location that the log file is saved to ?
2) with approx 3TB to be recovered (on a 4TB disk) (6455 files) does 3gb seem about the right size for a log file , as I could shrink the swap partition (currently 7.7gb) and expand the main partition to give more room,but this is only worth doing if I can be certain I will have enough room.
3) The 2nd option is to clone the USB boot drive over to a spare 120gb ssd (received another 2 port pcie card in the post this morning so can now have a spare port)
4) I assume that if I do a fresh install on the ssd then snapraid will complain that all the UUID's are different and the fix will fail ?
5) I could just run the fix without creating a log file , but would prefer to create one
in case of any errors.
I understand the risks involved with changing the partition sizes , but can take a fresh image (clonzilla) before doing anything.
any advice on which option to take would be appreciated .
Cheers, shady
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I realise that trying to guess the size of the fix.log is difficult because I assume it depends on both the number of files and the number of blocks that each file uses for parity.
I am really only posting this for info for others as I doubt I am the only one who did not realise how much free space is required for the fix.log. (it does mention in the manual that a few gigs are required , but I had either not noticed it or had forgotten about it when setting up this machine).
If others are using small USB sticks as boot devices then its better to be forwarned rather than run into issues of lack of space when attempting to do a recovery.
To avoid issues in the future I will either buy a sandisk extreme 32gb USB3 for the boot drive (I currently run Linux mint 17 mate) or use a 120gb ssd (seems a shame to waste the ssd on this server and I had plans for both the ssd and the spare sata port!)
I think Andrea is doing a great job on this project and the intention of any of my posts is to ask questions/provide real world results that might be useful to others.
Shady
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
having successfully merged my previous 2 snapraid machines into one new one with dual parity. a sync-H plus a check all reported no errors. On powering up the machine this morning one of the data disks failed to mount and had a corrupted partition table, I have rebuilt the table using gparted and started a Fix -l for just the dead disk (d3).
I then aborted this as I had not appreciated how much space would be required for the log file and dont think it will fit in the space available. It appears to be using approx 100mb for each 1% and with 3TB to be recovered will be very tight on the 3gb free space on my boot drive, which is only a 16gb USB3 sandisk Extreme (to save on sata ports)
My options/questions are
1) can I change the location that the log file is saved to ?
2) with approx 3TB to be recovered (on a 4TB disk) (6455 files) does 3gb seem about the right size for a log file , as I could shrink the swap partition (currently 7.7gb) and expand the main partition to give more room,but this is only worth doing if I can be certain I will have enough room.
3) The 2nd option is to clone the USB boot drive over to a spare 120gb ssd (received another 2 port pcie card in the post this morning so can now have a spare port)
4) I assume that if I do a fresh install on the ssd then snapraid will complain that all the UUID's are different and the fix will fail ?
5) I could just run the fix without creating a log file , but would prefer to create one
in case of any errors.
I understand the risks involved with changing the partition sizes , but can take a fresh image (clonzilla) before doing anything.
any advice on which option to take would be appreciated .
Cheers, shady
Hi Shady,
quick question on "log file": do you refer to the snapraid log file using option -l (--log) ?
Cheers,
Jens.
Yes I did mean the fix.log file that is located in /Home/(Yourusername)
I decided to reduce the swap partition and expand the normal partition of the USB stick so that I had 6gb free.
I then ran the following
sudo snapraid -d d3 -l fix.log fix (to rebuid my d3)
which gave the following positive results
17060071 processed in 12 Hours 23 Minutes
10816901 read/data errors
10816901 recovered/errors
Everything Ok
The total amount recovered was 2640.1 gig
and the fix.log file came to 5.2 gig
(I ended up with just under 700 meg free!!)
I realise that trying to guess the size of the fix.log is difficult because I assume it depends on both the number of files and the number of blocks that each file uses for parity.
I am really only posting this for info for others as I doubt I am the only one who did not realise how much free space is required for the fix.log. (it does mention in the manual that a few gigs are required , but I had either not noticed it or had forgotten about it when setting up this machine).
If others are using small USB sticks as boot devices then its better to be forwarned rather than run into issues of lack of space when attempting to do a recovery.
To avoid issues in the future I will either buy a sandisk extreme 32gb USB3 for the boot drive (I currently run Linux mint 17 mate) or use a 120gb ssd (seems a shame to waste the ssd on this server and I had plans for both the ssd and the spare sata port!)
I think Andrea is doing a great job on this project and the intention of any of my posts is to ask questions/provide real world results that might be useful to others.
Shady
Ahh, cool that's all is fine now!
Well, just specify a full path on "-l" log option and you can place it anywhere ;-)
or did I miss something?
e.g. snapraid -d d3 -l /mnt/logdevice/fix.log fix
Cheers, Jens.
I did wonder if it was as simple as that :)
Cheers,Shady