FreeNAS advice

[Mikensan]

Just a friendly reminder to those of you running FreeNAS - enable email alerts. Got the dreaded degraded email :-( Without the email I probably would've never really noticed (until much later anyway)

Quote

The volume bigboy state is DEGRADED: One or more devices are faulted in response to persistent errors. Sufficient replicas exist for the pool to continue functioning in a degraded state.

 

Little extra tid-bit... even spare drives can fail :-( Can't help but feel anger at seagate... within 1 year I had a disk fail - I replaced it with a HGST and RMA'd the dead disk. Upon receiving the new disk, I decided to keep it as a spare.. Well here we are another 2 years later and another seagate disk bites the dust. Phew I think, I have a spare. Pop the spare in... write errors galore after a couple hours of resilvering. 3 dead seagates within 3 years. Maybe I should've tested / burned-in the disk I got from the RMA, really wasn't expecting it to be dead. 

 

Thankfully I have a second separate volume for backups, but stinks to have $160x2 (originally paid) get flushed down the toilet.

 

/end rant

[Makkc]

Yeah, I don't rate Seagate Barracuda 3TB drives in particular having gone through 4 now!  Needless to say I steer well clear and haven't even tried their 'ironwolf' tech.  I lost a NAS array when a replacement drive did not successfully rebuild an array due to having errors.  Luckily it was a 4 bay RAID5 I think and got most the data back. I know backup first but, imho, quite often to back up your NAS you need a similar sized back up NAS especially if you want versioning.  Hmm. Well at the time I couldn't afford one and didn't really understand the difference in the setup between RAID disk fault tolerance and NAS Back up. Now I buy a new NAS and repurpose the old one to backup whilst steadily increasing the total TB storage space.  One day I'll retire and it will stop.

[Mikensan]

17 hours ago, Makkc said:

Yeah, I don't rate Seagate Barracuda 3TB drives in particular having gone through 4 now!  Needless to say I steer well clear and haven't even tried their 'ironwolf' tech.  I lost a NAS array when a replacement drive did not successfully rebuild an array due to having errors.  Luckily it was a 4 bay RAID5 I think and got most the data back. I know backup first but, imho, quite often to back up your NAS you need a similar sized back up NAS especially if you want versioning.  Hmm. Well at the time I couldn't afford one and didn't really understand the difference in the setup between RAID disk fault tolerance and NAS Back up. Now I buy a new NAS and repurpose the old one to backup whilst steadily increasing the total TB storage space.  One day I'll retire and it will stop.

Can't say my faith in barracudas is any good either. I try to be unbiased / impartial / objective - but when it hits home it's hard to be. Whole idea of raid is using inexpensive disks anyway right - all disks fail eventually. Just sucks when it's only been a handful of years.

 

ZFS snapshots are very lightweight, so versioning is pretty cheap. On a couple datasets I run snapshots every 15 minutes M-F from 6pm-5am, all day Sat/Sun. Retention is set for 2 weeks - so you can imagine how many versions I have available lol. 

 

You also should evaluate "critical" data - and at the very least have sufficient backup coverage. For me I have 5x 4TB for my main array, but only 3x3TB for backups. Majority of my data is media that can be recovered - unrecoverable data (pictures / documents / old schoolwork) probably only amounts to 1TB. 

[dalekphalm]

On 5/29/2018 at 3:41 PM, Mikensan said:

Just a friendly reminder to those of you running FreeNAS - enable email alerts. Got the dreaded degraded email :-( Without the email I probably would've never really noticed (until much later anyway)

 

Little extra tid-bit... even spare drives can fail :-( Can't help but feel anger at seagate... within 1 year I had a disk fail - I replaced it with a HGST and RMA'd the dead disk. Upon receiving the new disk, I decided to keep it as a spare.. Well here we are another 2 years later and another seagate disk bites the dust. Phew I think, I have a spare. Pop the spare in... write errors galore after a couple hours of resilvering. 3 dead seagates within 3 years. Maybe I should've tested / burned-in the disk I got from the RMA, really wasn't expecting it to be dead. 

 

Thankfully I have a second separate volume for backups, but stinks to have $160x2 (originally paid) get flushed down the toilet.

 

/end rant

I have had at least one failed Seagate HDD, along with a WD and some HGST ones too.

 

Honestly, for the vast majority of HDD's, it just comes down to luck.

 

I recommend burning in every HDD you get. New ones, refurbs, even RMA's.

 

My FreeNAS server currently has 5x Toshiba and 1x Seagate 3TB drives - it originally had 2x Seagate drives but one died after a few months and I replaced it with the 5th Toshiba.

 

The Seagate drives were well used though - I've had them for like 5+ years, so it dying after a few months in the FreeNAS server is not really a knock against Seagate.

[Mikensan]

1 minute ago, dalekphalm said:

I have had at least one failed Seagate HDD, along with a WD and some HGST ones too.

 

Honestly, for the vast majority of HDD's, it just comes down to luck.

 

I recommend burning in every HDD you get. New ones, refurbs, even RMA's.

 

My FreeNAS server currently has 5x Toshiba and 1x Seagate 3TB drives - it originally had 2x Seagate drives but one died after a few months and I replaced it with the 5th Toshiba.

 

The Seagate drives were well used though - I've had them for like 5+ years, so it dying after a few months in the FreeNAS server is not really a knock against Seagate.

Yea the logical half of my brain says this, just frustrating to have happen to. I've got drives that have lasted.. almost 18 years? So I've certainly cheated death you could say lol.

 

I also agree with burning in, might've avoided a lot of headache. I also bought these disks from a single vendor, and all their serial numbers were very close together. So I'm not entirely surprised 2 of the original 5 have died off. I planned for disks to fail, just sucks when they do.

[dalekphalm]

12 minutes ago, Mikensan said:

Yea the logical half of my brain says this, just frustrating to have happen to. I've got drives that have lasted.. almost 18 years? So I've certainly cheated death you could say lol.

 

I also agree with burning in, might've avoided a lot of headache. I also bought these disks from a single vendor, and all their serial numbers were very close together. So I'm not entirely surprised 2 of the original 5 have died off. I planned for disks to fail, just sucks when they do.

Yep, you can find some good guides online that show how to burn the drives in on FreeNAS using badblocks. It does mean you can't use the new drive until after you've burned it in though.

[Makkc]

Yup burning in can save future hangover especially as I totally agree the bad batch theory, these things are mass produced pretty quick cheap. and! you can get your money back so much easier if you find out sooner than later.

[Censored Tech]

Well.. I could say learn the terminal?

[dalekphalm]

11 hours ago, Censored Tech said:

Well.. I could say learn the terminal?

Terminal is useful for any Unix based OS, whether that's Linux, FreeBSD/FreeNAS, even macOS.

[Mikensan]

20 hours ago, Censored Tech said:

Well.. I could say learn the terminal?

Not sure how this would either save my disk or alert me when I'm away from said terminal? Point was to encourage people to configure alerting via email, when something is just working you never login - and thus would never know until maybe an update... Even in limp mode I was getting 80mbyte/s...

[Censored Tech]

i said that because free nas is based on free bsd