Andres Biront
Dabbler
- Joined
- Sep 3, 2013
- Messages
- 17
Helly everyone! I just wanted to share my experience and new build (not sure if it's the right subforum). And also, some scary dead drives photos (and not dead yet, in fact).
I had a 8 x 3TB disks, running in RAIDZ1 (4x3TB RAIDZ1 + 4x3TB RAIDZ1). I know, I know, it's not the best configuration, but I thought the chances of having 2 dead drives on the same array were thin, until in happened. It's my homelab and media, so there isn't much critical information, and I have local backups to external drives and then to the cloud for everything that is important to me.
Anyway, long story short, in 1 week I lost 4 drives. First one was quickly replaced for a 8TB drive (I was thinking to upgrade all of them in time... they served me well and were getting really old). It resilvered just fine, but two days later, another disk died. Instead of shutting everything down and wait for the replacement as I normally do, it was late and I thought "I'll get to it tomorrow".
The next morning another drive was dead (both with checksum and write errors), and the pool was faulted. Both drives were on the same array. Not a big deal, I had backups, but I really wanted to restore as a last option really...
Anyway, I proceeded to shut everything down, bought new disks, only 2 in fact, and power up hoping one of the 2 disks would come back from the dead. And it actually did, but another disk failed from the other array. That one with awful clicking noises, so it was actually quite dead :P
So far, 3 dead drives on 2 days... that caught my attention. I ended up replacing everything and installing other things that I bought to upgrade my NAS. So, my new build is:
Intel Xeon E5-2680
96GB DDR3 ECC
Some cheap Hunanzhi chinese board
Dell H310
8 x 8TB Toshiba X300 drives (don't blame me, I paid U$S80 each, new in box)
2 x 1.6TB Intel DC P3600 (MaxLBA 500GB, log and cache, running ~40 VMs with always sync. Paid U$S125 each, 85% life remaining)
2 x 10GbE (SFP+, broadcom based, don't remember which chipset. Dirt cheap, but I can saturate both links).
I'm now running RAIDZ2 :)
Finally, after looking at the old drives, I found 6 of them have signs of corrosion. I'll upload some pics to fuel nightmares. (Seagate Constellation ES.2)
Sorry for the low quality pics, they look even worse in real life.
I had a 8 x 3TB disks, running in RAIDZ1 (4x3TB RAIDZ1 + 4x3TB RAIDZ1). I know, I know, it's not the best configuration, but I thought the chances of having 2 dead drives on the same array were thin, until in happened. It's my homelab and media, so there isn't much critical information, and I have local backups to external drives and then to the cloud for everything that is important to me.
Anyway, long story short, in 1 week I lost 4 drives. First one was quickly replaced for a 8TB drive (I was thinking to upgrade all of them in time... they served me well and were getting really old). It resilvered just fine, but two days later, another disk died. Instead of shutting everything down and wait for the replacement as I normally do, it was late and I thought "I'll get to it tomorrow".
The next morning another drive was dead (both with checksum and write errors), and the pool was faulted. Both drives were on the same array. Not a big deal, I had backups, but I really wanted to restore as a last option really...
Anyway, I proceeded to shut everything down, bought new disks, only 2 in fact, and power up hoping one of the 2 disks would come back from the dead. And it actually did, but another disk failed from the other array. That one with awful clicking noises, so it was actually quite dead :P
So far, 3 dead drives on 2 days... that caught my attention. I ended up replacing everything and installing other things that I bought to upgrade my NAS. So, my new build is:
Intel Xeon E5-2680
96GB DDR3 ECC
Some cheap Hunanzhi chinese board
Dell H310
8 x 8TB Toshiba X300 drives (don't blame me, I paid U$S80 each, new in box)
2 x 1.6TB Intel DC P3600 (MaxLBA 500GB, log and cache, running ~40 VMs with always sync. Paid U$S125 each, 85% life remaining)
2 x 10GbE (SFP+, broadcom based, don't remember which chipset. Dirt cheap, but I can saturate both links).
I'm now running RAIDZ2 :)
Finally, after looking at the old drives, I found 6 of them have signs of corrosion. I'll upload some pics to fuel nightmares. (Seagate Constellation ES.2)
Sorry for the low quality pics, they look even worse in real life.
