ZFS Uses:
ZFS is really great for backups. We all know that the best backup is two
copies. My previous company had a product strategy that allows incremental
backups backups "forever" and always resulted in a "full."
Configure your "master" or server using ZFS.
Configure your replication target a
Rich Pieri wrote:
> On Wed, 27 Jul 2022 16:13:31 -0400
> ma...@mohawksoft.com wrote:
>
> > A stripe over two mirrors is not as reliable as RAID6. If you have 4
> > drives arranged in two mirrors, each mirror can only survive the loss
> > of one drive. So, your system, if it loses 2 drives, has a
On Wed, 27 Jul 2022 16:13:31 -0400
ma...@mohawksoft.com wrote:
> A stripe over two mirrors is not as reliable as RAID6. If you have 4
> drives arranged in two mirrors, each mirror can only survive the loss
> of one drive. So, your system, if it loses 2 drives, has a 33.33%
> chance of losing half
I know this is OT, but I removed myself from the BLU memberships, but
somehow I still get some emails that must be from a separate email
list. Please remove me.
Jerry Natowitz
j.natow...@gmail.com
On 7/27/2022 4:17 PM, Rich Pieri wrote:
On Wed, 27 Jul 2022 15:56:41 -0400
ma...@mohawksoft.com
On Wed, 27 Jul 2022 15:56:41 -0400
ma...@mohawksoft.com wrote:
> RAID has that reputation, but have you measured it recently? The ghz,
> multicore beasts we have in our computers these days, with specialized
> cryptography instructions are really very good.
Yes, and the server I have today could
> On Wed, 27 Jul 2022 15:06:50 -0400
> ma...@mohawksoft.com wrote:
>
>> Why two mirrors and not a RAID5 - or is RAID6 more appropriate 4T?
>
> Because parity RAID (RAIDZ-1/2/3) is slow.
A stripe over two mirrors is not as reliable as RAID6. If you have 4
drives arranged in two mirrors, each mirror
> On Wed, 27 Jul 2022 15:06:50 -0400
> ma...@mohawksoft.com wrote:
>
>> Why two mirrors and not a RAID5 - or is RAID6 more appropriate 4T?
>
> Because parity RAID (RAIDZ-1/2/3) is slow.
RAID has that reputation, but have you measured it recently? The ghz,
multicore beasts we have in our computers
On Wed, 27 Jul 2022 15:06:50 -0400
ma...@mohawksoft.com wrote:
> Why two mirrors and not a RAID5 - or is RAID6 more appropriate 4T?
Because parity RAID (RAIDZ-1/2/3) is slow.
--
\m/ (--) \m/
___
Discuss mailing list
Discuss@lists.blu.org
http://lists.
ma...@mohawksoft.com wrote:
> Gotta ask again. Why mirror and not "RAID?"
The tradeoffs are:
read IOPS.
read xfer
write IOPS
write xfer
capacity loss
recovery after 1 disk lost
recovery after 2 disks lost
If you have four disks, your choic
Gotta ask again. Why mirror and not "RAID?"
> House non-media server:
>
> pool: home
> state: ONLINE
> scan: scrub repaired 0B in 00:24:11 with 0 errors on Thu Jul
> 21 02:24:12 2022
> config:
>
> NAME STATE READ WRITE
> CKSUM
> home
Why two mirrors and not a RAID5 - or is RAID6 more appropriate 4T?
> On Wed, 27 Jul 2022 11:51:58 -0400
> ma...@mohawksoft.com wrote:
>
>> I just replaced a bad drive with sdk. I've had this configuration
>> since 2T hard disks became afordable. It's not getting on in years,
>> and I'm only gett
House non-media server:
pool: home
state: ONLINE
scan: scrub repaired 0B in 00:24:11 with 0 errors on Thu Jul
21 02:24:12 2022
config:
NAME STATE READ WRITE CKSUM
home ONLINE 0 0 0
On Wed, 27 Jul 2022 11:51:58 -0400
ma...@mohawksoft.com wrote:
> I just replaced a bad drive with sdk. I've had this configuration
> since 2T hard disks became afordable. It's not getting on in years,
> and I'm only getting 4T, but it is redundant. The logs and cache
> drives are ssd.
root@marlle
> On 7/26/22 19:43, ma...@mohawksoft.com wrote:
>> Now, the exciting part!!! ZFS does a lot of awesome things.
>
> For years I was a big fan of Linux software raid 1 (could never afford
> fancier raid levels), until disks outgrew it: In my raid 1 setups one
> hiccough and one whole damn disk needs
On 7/26/22 19:43, ma...@mohawksoft.com wrote:
Now, the exciting part!!! ZFS does a lot of awesome things.
For years I was a big fan of Linux software raid 1 (could never afford
fancier raid levels), until disks outgrew it: In my raid 1 setups one
hiccough and one whole damn disk needs to be r
How many members are fairly knowledgeable about ZFS?
How many are curious but have little exposure?
How many have no idea what ZFS is or why?
I have mixed feeling about ZFS. Its amazing, everything it does and the
way it manages storage. The model is almost perfect. IMHO, of course.
I have had to
On 7/26/22 15:12, Rich Pieri wrote:
If you are getting ~100MB/s sustained then that is about as fast as
you can go on spinning disks regardless of the bus.
Before I threw in the complication of copies=2, I think was writing
something in the 90s MBs/s. Which I thought was very nice. For $130 (
On Tue, 26 Jul 2022 20:09:02 -0400
Dan Ritter wrote:
> You will be surprised, as I recently was, that the new
> enterprise spinning disks from Seagate and WD boast claimed
> 280MB/s transfers, which can actually be sustained at
> 200-250MB/s depending on the model. IOPS has increased, too,
> thou
Rich Pieri wrote:
> If you are getting ~100MB/s sustained then that is about as fast as you
> can go on spinning disks regardless of the bus.
You will be surprised, as I recently was, that the new
enterprise spinning disks from Seagate and WD boast claimed
280MB/s transfers, which can actually be
On Tue, 26 Jul 2022 14:41:55 -0700
Kent Borg wrote:
> It took hours. The fact that ZFS assumes that numerically distant
> LBAs will be physically distant, to the extent that assumption holds,
> will make for a lot of extra head movement. Maybe Linux can reorder
> writes around that a little, but
On 7/25/22 08:58, Rich Pieri wrote:
zfs set copies=2 extern_backup_pool
Of my two new ZFS disks, I think have converted one of them over to "copies=2"
with the data all in double copies. And that took quite some time!
As fast as USB-C is, as fast as modern spinning rust is, copying aroun
On Tue, 26 Jul 2022 12:54:43 -0400
ma...@mohawksoft.com wrote:
> That's interesting. What version of Linux? I'm using Ubuntu, but at
> work we were using CentOS7.
Debian. I switched from eSATA to USB 3 with Debian I think 8 but it
could have been 9. Definitely been using USB with Debian 10 and 11
Something to test, perhaps?
https://www.cnx-software.com/2020/08/12/how-to-fix-unreliable-usb-hard-drives-stalled-transfers-linux-windows/
> On Tue, 26 Jul 2022 11:10:36 -0400
> ma...@mohawksoft.com wrote:
>
>> > I have been running backups to ZFS on USB storage for years. Some
>> > zfs send/recei
That's interesting. What version of Linux? I'm using Ubuntu, but at work
we were using CentOS7.
> On Tue, 26 Jul 2022 11:10:36 -0400
> ma...@mohawksoft.com wrote:
>
>> > I have been running backups to ZFS on USB storage for years. Some
>> > zfs send/receive with data copies=2, some rsync with mir
On Tue, 26 Jul 2022 11:10:36 -0400
ma...@mohawksoft.com wrote:
> > I have been running backups to ZFS on USB storage for years. Some
> > zfs send/receive with data copies=2, some rsync with mirrored USB
> > drives. Never experienced any problems of this sort.
>
> USB-3 or USB-2?
USB 3. USB 2 w
> On Tue, 26 Jul 2022 08:55:49 -0400
> ma...@mohawksoft.com wrote:
>
>> Be that as it may, its been my experience (anecdotal) that the Linux
>> USB storage stack has issues and tends to go offline under highly
>> concurrent heavy load. This faults the drive and fails ZFS, YMMV
>
> I have been runni
On Tue, 26 Jul 2022 08:55:49 -0400
ma...@mohawksoft.com wrote:
> Be that as it may, its been my experience (anecdotal) that the Linux
> USB storage stack has issues and tends to go offline under highly
> concurrent heavy load. This faults the drive and fails ZFS, YMMV
I have been running backups
I've been using ZFS for about 10 years personally and in my company's
product. I love the way ZFS presents storage. I have issues with ZFS's
internals. Its kind of ugly.
Be that as it may, its been my experience (anecdotal) that the Linux USB
storage stack has issues and tends to go offline under
On 7/25/22 11:53, Dan Ritter wrote:
[lots of interesting stuff]
Stuff that I mostly can't play with, yet, because I am waiting for the
first tranche of my copying data from one of my disks to itself (to
convert from one-copy data to two-copy data), takes a *bleeping* long time.
Thanks, how
Kent Borg wrote:
> zpool create extern_backup_pool /dev/sda
> zpool export extern_backup_pool
zpool export prepares a pool to be used on a different machine -
for an external portable disk, you might do this in order to
unmount it.
You definitely don't need to do that right after you create
On 7/25/22 09:26, Kent Borg wrote:
Question: Can I still do that, on disks already in use??
Oops, you already answered my question.
Thanks,
-kb
___
Discuss mailing list
Discuss@lists.blu.org
http://lists.blu.org/mailman/listinfo/discuss
On 7/25/22 08:58, Rich Pieri wrote:
I recommend enabling
copies=2 if you don't mirror your backup volumes:
zfs set copies=2 extern_backup_pool
I can do that on a single disk? Interesting!
Question: Can I still do that, on disks already in use??
-kb
___
On Mon, 25 Jul 2022 08:34:04 -0700
Kent Borg wrote:
> To check whether the data all reads back correctly do a scrub (takes
> a long time, and note the scrub happens on the underlying pool not
> the dataset,
> the volume does not have to be mounted to do a scrub, though the pool
> needs to be imp
I recently had a disk I use for external backup lose its mind. And
though XFS has been mostly good to me for many years, I decided it was
time to try ZFS. The fact it will notice any data errors it encounters,
and can check the entire disk for errors, is reassuring. And the fact it
is designed
34 matches
Mail list logo
