FreeNAS extremely slow

[Stux]

Maybe this will help too: https://forums.freenas.org/index.php?threads/comprehensive-diagram-of-the-zfs-structure.38865/ ;)

There can be more than one pool ;)

 

[MR. T.]

You can partition the 3TB drive into a 2TB and 1TB partition... and then make a 2TB mirror and a 1TB mirror.

Maybe not so useful.

But perhaps it would make sense to treat your 3 2TB drives as 6 1TB partitions and combine with your 1TB drive for a 7 wide RAIDZ2. You could then lose any 2 1TB partitions... or even the whole 2TB drive.

I had no idea i could break apart a disk into partitions and use them separately. this might be useful for a pool that doesn't hold important data. On the above example losing a 2tb disk would put me at risk but wouldn't kill off the entire pool... i might consider that for data that can be recovered without much effort.

Does freeNAS allows to do this over the UI? or do i have to go into the console?

This should be put in bold.

Indeed. Even i who am a huge noob can see immediately that there are substantial risks and drawbacks involved (like having the disk seeking for data on both partitions all the time).

 

[Bidule0hm]

There can be more than one pool ;)

Would be a bit tight to fit multiples pools... :p

 

[Stux]

I had no idea i could break apart a disk into partitions and use them separately. this might be useful for a pool that doesn't hold important data. On the above example losing a 2tb disk would put me at risk but wouldn't kill off the entire pool... i might consider that for data that can be recovered without much effort.

Does freeNAS allows to do this over the UI? or do i have to go into the console?

Definitely command line kung-fu time.

I use this technique a few times when doing in-place upgrades from one pool to another... ie to go from say... 3 drive RAIDZ1 to 6 drive RAIDZ2... ie the 3 extra drives... get partitioned into 6 partitions (slices?)... then you can make a 6 way z2... copy the data... then upgrade the drives by replacing the partitions with whole drives. If you do it right, you have minimimal risk. It just takes ages to resilver each drive.

And fundamentally, this is how things like a Drobo deal with disparate disk sizes in raids... Unfortunately, since you can't add/remove drives from a vdev, you can't get the same functionality as a Drobo has.

Indeed. Even i who am a huge noob can see immediately that there are substantial risks and drawbacks involved (like having the disk seeking for data on both partitions all the time).

 

[Stux]

I use this technique a few times when doing in-place upgrades from one pool to another...

For an extreme example of the shenanigans, see the ... Lets Dance...

 

[MR. T.]

I use this technique a few times when doing in-place upgrades from one pool to another... ie to go from say... 3 drive RAIDZ1 to 6 drive RAIDZ2... ie the 3 extra drives... get partitioned into 6 partitions (slices?)... then you can make a 6 way z2... copy the data... then upgrade the drives by replacing the partitions with whole drives. If you do it right, you have minimimal risk. It just takes ages to resilver each drive.

Mind Blown!
Knowing this at the beginning of the year would have saved me a lot of grief.
I could have created the pool of a single disk and slowly replaced the partitions by proper disks. Instead i had to go and find all the disks i had lying around (many of them really crappy and erroring all over the place) so i could have the 11 disk wide pool i wanted.

It was these rubbish disks that spawned this thread in the first place.

All i need to know now is how to do a proper burn in.
how many times should i read/write the disks? or should i just do it for a certain amount of time?

If a disk is having problems (SMART errors) is it possible that stressing the disk (burn in) would make it mark the bad sectors as bad and then work fine from then onwards? or as soon as it starts might as well bin it and buy a new one?

 

[Stux]

Mind Blown!
Knowing this at the beginning of the year would have saved me a lot of grief.
I could have created the pool of a single disk and slowly replaced the partitions by proper disks. Instead i had to go and find all the disks i had lying around (many of them really crappy and erroring all over the place) so i could have the 11 disk wide pool i wanted.

It was these rubbish disks that spawned this thread in the first place.

Sorry :)

All i need to know now is how to do a proper burn in.

https://forums.freenas.org/index.php?threads/how-to-hard-drive-burn-in-testing.21451/

how many times should i read/write the disks? or should i just do it for a certain amount of time?

The above burnin process terminates.

It takes 3 days or so for a set of 4TB drives.

If a disk is having problems (SMART errors) is it possible that stressing the disk (burn in) would make it mark the bad sectors as bad and then work fine from then onwards? or as soon as it starts might as well bin it and buy a new one?

Yes... depending on what exactly is the problem. For example an old disk might just have a few blocks it can't currently read, a "pending sector". There are techniques for "repairing" pending sectors etc...

Do the research etc AFTER you come across that problem ;)

 

[MR. T.]

The above burnin process terminates.

It takes 3 days or so for a set of 4TB drives.

I had read that post but thought the process thad to be repeated X amount of times.
Running it once makes it simpler.

Yes... depending on what exactly is the problem. For example an old disk might just have a few blocks it can't currently read, a "pending sector". There are techniques for "repairing" pending sectors etc...

Do the research etc AFTER you come across that problem ;)

I have so many drives chugging SMART errors that is crazy. So far i have been removing the drives but if i can salvage them for non critical data... would be nice.


I have one more question:
My motherboard has 12 sata ports, the LSI LBA i bought has 16 but doesn't seem to recognise 8tb disks.
Of those ports i might have 6 ports left.
I cant get another LBA because the motherboard only has 1 pci express slot.
Any simple (and cheap) solution for this?

If port multipliers are the way to go is there some i should get/ should avoid?

 

[Stux]

You can use a SAS expander with the HBA.

Not sure about your 8TB issue.

 

[MR. T.]

Not sure about your 8TB issue.

It's probably because is an old model... probably doesn't support it.
It's not problematic at this point because i "only" need 11 8tb disks and have enough ports on the motherboard for that

You can use a SAS expander with the HBA.

I googled SAS expander and found thousands of hits for some boards but all of them use PCI express and i have no more slots.

I also found a few SAS expander enclosures... but they are expensive.
...very expensive.










...Oh damn they are expensive!

 

[Stux]

It's probably because is an old model... probably doesn't support it.
It's not problematic at this point because i "only" need 11 8tb disks and have enough ports on the motherboard for that


I googled SAS expander and found thousands of hits for some boards but all of them use PCI express and i have no more slots.

They lie. SAS expanders only use PCIe for power and positioning. Do some searching on the forum for solutions

Like: http://www.ebay.com.au/itm/like/231...5960-0%26rvr_id%3D1113947638206&ul_noapp=true

I also found a few SAS expander enclosures... but they are expensive.
...very expensive.










...Oh damn they are expensive!