A friend gave me an oldish HP Elitedesk. I'm going to pull the i7-4770 out and throw it in a SUPERMICRO MBD-C7Z87-O ATX Motherboard LGA 1150 Intel Z87 DDR3 1600. I've also ordered 32gb of memory(max). I've got 4 drives from 2 old Buffalo Nas enclosures I'll be using to start. ATX case with a rosewell 4 bay hot swap cage. Good start? Any gotchas I should be aware of?
-
Important Announcement for the TrueNAS Community.
The TrueNAS Community has now been moved. This forum has become READ-ONLY for historical purposes. Please feel free to join us on the new TrueNAS Community Forums
First Build
- Thread starter reppard
- Start date
- Joined
- May 17, 2014
- Messages
- 3,611
What is your boot device?
Some new people to TrueNAS don't know that TrueNAS uses a separate device for booting. Not saying you don't, just looking at your post count and what you've already written.
Next, one of the most important things for us to know, is your use case(s). Media storage, backups, etc...?
With that, the pool layout can follow.
Last, what is the brand, model and size of your disks?
What layout do you plan for your pool, and why?
Again some new users don't know enough about ZFS to plan on how they want to lay out the disks. For example, larger disks like >=2TB suggest using either Mirrors or RAID-Z2 because RAID-Zx re-sync with larger disks takes time. And a second failure, even a single block, can impact the resilver, (aka disk re-sync after replacement).
Some new people to TrueNAS don't know that TrueNAS uses a separate device for booting. Not saying you don't, just looking at your post count and what you've already written.
Next, one of the most important things for us to know, is your use case(s). Media storage, backups, etc...?
With that, the pool layout can follow.
Last, what is the brand, model and size of your disks?
What layout do you plan for your pool, and why?
Again some new users don't know enough about ZFS to plan on how they want to lay out the disks. For example, larger disks like >=2TB suggest using either Mirrors or RAID-Z2 because RAID-Zx re-sync with larger disks takes time. And a second failure, even a single block, can impact the resilver, (aka disk re-sync after replacement).
I've got a few smaller SSDs I'm considering, but to keep all the 6xSata ports on the board free for drives, I've also considered getting some usb3 thumb drives to build the boot pool.
My use case(s) will be very varied. Both media storage and backups for my work station. I plan on throwing nginx in a jail to serve up a few static sites and I'll try to get docker running in a VM to serve up some other services.
Basically I'm working towards decommissioning an old Macmini I've been using for a catch all server for years. It's been running linux with a software raid on some usb HDs. Obviously not the best option but its what I had to work with.
Currently I have 2x2TB WD and 2x4TB WD (blue and green I believe). I was planning on just creating a mirrored pool for each set with plans to add a pool of very large drives in the future(budget permitting).
I was curious if something like this would work:
MIRROR:
STRIPE: 2TB + 4TB
STRIPE: 2TB + 4TB
The idea being to maximize capacity and then extend with larger drives in the future? I don't even know if this is possible and what performance would look like even if it was. Thanks for post and any future feedback.
- Joined
- May 17, 2014
- Messages
- 3,611
Yes, you can have all 4 disks in the same pool, with the same sized disks mirroring each other. The mirrored pairs are striped, though because the are uneven sizes, ZFS will write more to the 4TB pair to maintain same amount of free space on them. Best case, performance is better with 2 x 2 disk mirrors than 1 x 2 disk mirror.
This allows you at some future date to replace the smaller mirror pair with larger, to grow the pool. This is done live, by replacing 1 disk at a time. When done, replace the other. For safeties sake, if you have a spare disk port & bay, you can replace in place, meaning you tell ZFS to replace an existing disk with another disk while both are in the system. This causes ZFS to make a temporary 3 way mirror and when the re-silver is complete, break off the old disk, restoring the mirror back to 2. This is mostly automatic, except installing the new disk, issuing the replacement command, and removing the old disk.
Or you can add a new mirrored pair to your pool, (so 3 x 2 disk Mirrors). Or create a new ZFS pool. Depends on what you need.
WD Blue & Green likely are not suitable for use with ZFS, if they have the TLER set high, or Idle Head Parking enabled. TLER is good for desktop drives, not good for ZFS or other RAID configurations. Basically TLER, (Time Limited Error Recovery, Seagate has similar feature, but different name for it), means on read error, the drive will give up error recovery sooner, like 7 seconds. But desktop drives generally default to more than 1 minute.
With short TLER, ZFS finds an error, reads the redundant block from the Mirror disk, supplies it to the user and re-writes the good data on to the bad block. SATA will automatically spare out a bad block on write. Since ZFS wrote good data automatically, you get your redundancy back quickly.
However, with long TLER, like a minute, ZFS, (and some other RAID software or firmware), may decide that the disk that is no longer responding, is completely bad. The disk may be dropped off the Mirror as faulted, instead of a simple bad block. So in this case, you loose your redundancy completely.
Next Idle Head Parking can cause the disks to be not ready with ZFS wants to do something. This can cause errors, though sometimes harmless errors.
With ZFS, if you loose a vDev, (aka both disks in a mirrored pair), completely, you loose everything. Full restore from backup time. But, loose a block on 1 disk, if their are spare blocks on that disk, ZFS will auto-correct.
This allows you at some future date to replace the smaller mirror pair with larger, to grow the pool. This is done live, by replacing 1 disk at a time. When done, replace the other. For safeties sake, if you have a spare disk port & bay, you can replace in place, meaning you tell ZFS to replace an existing disk with another disk while both are in the system. This causes ZFS to make a temporary 3 way mirror and when the re-silver is complete, break off the old disk, restoring the mirror back to 2. This is mostly automatic, except installing the new disk, issuing the replacement command, and removing the old disk.
Or you can add a new mirrored pair to your pool, (so 3 x 2 disk Mirrors). Or create a new ZFS pool. Depends on what you need.
WD Blue & Green likely are not suitable for use with ZFS, if they have the TLER set high, or Idle Head Parking enabled. TLER is good for desktop drives, not good for ZFS or other RAID configurations. Basically TLER, (Time Limited Error Recovery, Seagate has similar feature, but different name for it), means on read error, the drive will give up error recovery sooner, like 7 seconds. But desktop drives generally default to more than 1 minute.
With short TLER, ZFS finds an error, reads the redundant block from the Mirror disk, supplies it to the user and re-writes the good data on to the bad block. SATA will automatically spare out a bad block on write. Since ZFS wrote good data automatically, you get your redundancy back quickly.
However, with long TLER, like a minute, ZFS, (and some other RAID software or firmware), may decide that the disk that is no longer responding, is completely bad. The disk may be dropped off the Mirror as faulted, instead of a simple bad block. So in this case, you loose your redundancy completely.
Next Idle Head Parking can cause the disks to be not ready with ZFS wants to do something. This can cause errors, though sometimes harmless errors.
With ZFS, if you loose a vDev, (aka both disks in a mirrored pair), completely, you loose everything. Full restore from backup time. But, loose a block on 1 disk, if their are spare blocks on that disk, ZFS will auto-correct.
jgreco
Resident Grinch
- Joined
- May 29, 2011
- Messages
- 18,680
Some additional information on TLER can be found here:
www.truenas.com
Checking for TLER, ERC, etc. support on a drive
This resource was originally created by user: jgreco on the TrueNAS Community Forums Archive. Please DM this account or comment in this thread to claim it. One of the problems with consumer-grade hard drives is that most of them will hang in the event that they run into an error, and will...
www.truenas.com
