samba+zfs

[Kurt Touet]

On Wed, Nov 9, 2011 at 1:07 AM, Daniel O'Connor <doconnor at gsoft.com.au> wrote:
>
> On 09/11/2011, at 17:32, Garrett Cooper wrote
>>> dd's of large files (spooled backups going to tape) to /dev/null are as slow as Samba.
>>
>>    - Dedupe?
>
> Nope.
>
>>    - Compression?
>
> On the mail spool & ports, but not on the tape spool.
>
>>    - How much RAM?
>
> 8GB.
>
>>    - What debug options do you have enabled in the kernel?
>
> It is 8.2-GENERIC so.. no WITNESS (for example)
>
>>    I've been noticing a slowdown in some respects with NFS/SMB, but I
>> suspected it was because I have an re(4) based NIC. ZFS has also wired
>> down a lot of my system memory for the L2ARC…
>
>
> re isn't great but I wouldn't expect it to slow down over time.. Unless bounce buffers got used more and more or something.
>
> I have an em0 card in this system - but in any case it is slow locally (i.e. dd a large file with 64k block size).
>
> --
> Daniel O'Connor software and network engineer
> for Genesis Software - http://www.gsoft.com.au
> "The nice thing about standards is that there
> are so many of them to choose from."
>  -- Andrew Tanenbaum
> GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C
>
>

Right now (while experience slow writes via samba+zfs) this is general
read speed off a 4 x 1.5TB sata2 raidz1:

# dd if=test.file of=/dev/null
13753502+1 records in
13753502+1 records out
7041793036 bytes transferred in 100.020897 secs (70403218 bytes/sec)

That's not in the same ball park of slow writes, but it is below what
I expect for reads.

My setup is a little odd:  4x1.5tb raidz sata2 on mobo + 2 x 2tb
mirror on sata1 pci controller, zfs v28, stable/9 r227357, amd x4 810
2.6ghz, 4gb ram, no dedupe, no compression, daily snapshots saved for
7 days

The above file read was stored before the 2 x 2tb mirror addition, so
it was a solely read off the sata2 mobo ports.   Reading off of
something more recent (and split amongst both raidz1 and mirror
vdevs):

# dd if=test2.file of=/dev/null
9154715+1 records in
9154715+1 records out
4687214153 bytes transferred in 82.963181 secs (56497522 bytes/sec)

This is, again, seems slower than usual, but not as terrible as the
write speeds that I've been seeing via samba.