Sort by time | Sort by thread (beta)
Link to this comment:
Share to:
Loading...
- Your queue is empty. Add videos to your queue using this button:
or sign in to load a different list.
Loading...Saving...
couldn't you just unplug the drive?
thegenrl 1 year ago 3
raid-z2 equivalent raid-6 except there is no "write holes"
recha36 1 year ago
How is this any different from RAID-6?
NineInchSamusAran 2 years ago
Copy on write style data writes. Huge performance increase over traditional Raid-6. Raid-6 only adds an extra parity. (If I remember correctly)
SteveCrothers 1 year ago
It appears that ZFS didn't recover after each drive failure until he unplugged the failed drive? Or was it coincidence that he unplugged the drive just as ZFS started recovering?
cyberkreiger 3 years ago 7
Yep. its a bug in solaris. BUt if you try and tell a sun person that, they get really pissy.
mickrussom 3 years ago 4
What's the likelihood that a hard drive would implode and short out connectors and/or break traces to create shorts?
Totally acceptable circumstances.
DiskoDude 3 years ago
I have a test system running Opensolaris 2008.11 with 8 disks in a raidz2 pool. To test this, I unplugged the power on two of the disks at once(Unlike Sun's marketing dept., we can't justify the destruction of a disk here). The reads paused for about 5 seconds and then resumed without any intervention.
It appears that this bug was fixed between April and November 2008.
kingsal13 2 years ago 7
Can I just plug in a new drive that's bigger and then have the system even out and also utilize the extra space gained?
bjeah 2 years ago
If you have 3 x 500GB drives and 1 x 640GB drive, you're missing out on ~140GB of data. It uses only the size of the smallest vdev. However, if you replace every 500GB drive with a 640GB one at a time, it'll resync each disk and then utilise the extra space once all drives are 640GB.
kashzone 2 years ago
Is this bug 6667208 you are talking about?
capitalistpig1968 2 years ago
hi,
how can i get this video???
katilemre 3 years ago
ya try that with a linux server!
Ive done that with windows but never got quite the same result.
rabbitnightmare 3 years ago
there is no ZFS in linux. Only in Solaris and FreeBSD
analogfantasy 2 years ago
That isn't correct. There is an implementation using FUSE. This isn't optimal ofcourse, so just as you are probably looking for an in-kernel driver, I am too :)
WretchedWinston 2 years ago
FUSE isn't really official is it? I was talking about official support
analogfantasy 2 years ago
And now in NetBSD
herdware 2 years ago
Comment removed
z0sasha 3 years ago 2
The difference is that this is commodity hardware, no EMC, no FC, no raid card, no added software. All that is needed comes with OpenSolaris for free.
jphughesatsun 3 years ago
Raid 6 comes with Linux for free, runs on commodity hardware, no EMC, no FC, no raid card, no added software, et.c...
ZFS has lots of other things going for it though, but those weren't demonstrated in this video.
cyberkreiger 3 years ago 3
Just more system load per unit of IO, bugs, cant make vdevs part of vdevs.
ZFS is cool, for sure, but its not a be all end all by any stretch. I've been able to break it in Solaris 10u5 where UFS would not break under the same stresses.
mickrussom 3 years ago
what io meter utility is that on the left?
NetSyphon 3 years ago