Questions about Drives - bad sectors, defrag, replacement etc.

[zenorb]

Hi

Building my first FreeNAS and need help going from zero 2 hero.

Looking at 14 x 3TB drive internal plus 8 x 3TB external via eSATA (if OK - see other post) for a total of 22 drives.

I'm assuming I should go RAIDZ2 over RAIDZ at least.

Does FreeNAS self defrag drives and check for bad sectors and identify, notify and repair?

I see from the FAQ that the best way to identify a failed drive is to note it's serial number. It would seem that I should write down all serial numbers of drives when building system and keep this document so I know which one to pull. Any other tricks or am I on the right path?

It appears all drives need to be the same size or else smallest drive determines capacity of all drives - is this right?

I was looking at adding external 4 bay enclosures via eSATA in RAID 5 for additional redundancy but now assume this would be pointless as they need to be JBODs.

How do I get notified when a drive goes bad - do I get emailed?

I looked under 4.1 Reporting and 9.6 Alert in 8.0.3 Guide cursory glance but couldn't see where this is mentioned and may have missed it.

For RAIDZ2 is capacity roughly n-2?
(so 22 x 3TB = 66 - 6 = 60TB)

Cheers

 

[cyberjock]

You did ZERO reading about ZFS if you are discussing defrag with RAIDZ2 or RAIDZ.

Like I said in the other thread, and you have made pretty clear, you want to go from zero to hero with no work. It doesn't work that way. Start reading and come back when you have actual questions to ask without expecting the forum to handhold you.

RTFM, RTFF, read my guide, and just do some searching and maybe some experimenting. If you don't do these things and rush into it you'll be a hero for a few days until you realize you screwed up and lost all your data. Then you'll be a zero again.

 

[zenorb]

Actually I did a lot of reading about ZFS over a number of days but missed the explicit part on defrag. Incorrect assumption on your part. Fortunately I've just worked out how to unsubscribe from all my posts so I don't have to read these answers any more.

 

[Stephens]

I'm assuming I should go RAIDZ2 over RAIDZ at least.

I would.

Does FreeNAS self defrag drives and check for bad sectors and identify, notify and repair?

ZFS Defrag
FreeNAS Scrub

I see from the FAQ that the best way to identify a failed drive is to note it's serial number. It would seem that I should write down all serial numbers of drives when building system and keep this document so I know which one to pull. Any other tricks or am I on the right path?

You're on the right path.

It appears all drives need to be the same size or else smallest drive determines capacity of all drives - is this right?

Correct.

I was looking at adding external 4 bay enclosures via eSATA in RAID 5 for additional redundancy but now assume this would be pointless as they need to be JBODs.

I don't really understand what you're trying to do/accomplish here.

How do I get notified when a drive goes bad - do I get emailed?

If any user of FreeNAS made it this far into your questions, they'd probably be frustrated at this point. Pretty much all FreeNAS documentation covers this pretty clearly. Yes, if you set up e-mail notifications, you will receive an e-mail status report. Of course that report is scheduled for 3:00 a.m., and you want to have your system off at that time to save power, so *shrug*.

I looked under 4.1 Reporting and 9.6 Alert in 8.0.3 Guide cursory glance but couldn't see where this is mentioned and may have missed it.

Full on frustration mode. You know there's a section which probably answers your questions, but rather than read it, you'd rather ask and have others answer what's already documented? I'm not going to snap on you, but I will ask that you be fair about the treatment you've received here. Shouldn't you have respect for other people's time? You're effectively saying your time/energy is more imporant than everyone else's. Sure they don't have to answer your questions, but they read them (time) with the expectation of trying to help (mental energy). Then they get to the end and see it's a bunch of basic stuff that's already documented. Here's how it comes off... "I know there's a sign there that says no parking, but I just wanted to know if I can park there." Come on. Surely you can see how that'd get at least a little frustrating for the traffic cop standing there. Meet us half way.

For RAIDZ2 is capacity roughly n-2? (so 22 x 3TB = 66 - 6 = 60TB)

My god. Please don't do this. At least not until you've done more reading on optimal zpool/vdev sizes. A 22-drive vdev would take forever to rebuild if you have to replace a drive.

You didn't really mention what type of data you plan to store. Do you really need it all in one NAS? What if your power supply goes flaky and takes out all 22 of your drives? Is your data critical? Will you have backups? RAID isn't backup. If you have your data spread across multiple NAS's, if one goes completely belly up, you don't lose what's on the other. That may or may not be feasible in your situation.