Re: ZFS status now in June? / stable enough for a file server?

From: Bernd Walter <ticso_at_cicely12.cicely.de>
Date: Sat, 23 Jun 2007 21:14:20 +0200
On Sat, Jun 23, 2007 at 10:22:10PM +0400, Dmitry Morozovsky wrote:
> On Fri, 22 Jun 2007, Bernd Walter wrote:
> 
> BW> Yours is fat as well:
> BW> CPU: AMD-K6tm w/ multimedia extensions (167.05-MHz 586-class CPU)
> BW>   Origin = "AuthenticAMD"  Id = 0x561  Stepping = 1
> BW>   Features=0x8001bf<FPU,VME,DE,PSE,TSC,MSR,MCE,CX8,MMX>
> BW>   AMD Features=0x400<<b10>>
> BW> real memory  = 167772160 (160 MB)
> BW> avail memory = 158793728 (151 MB)
> BW> 
> BW> Many things tuned to get ZFS runnning.
> BW> But it is used just as backup another bigger ZFS Server with
> BW> zfs receive.
> BW> I'm not shure if it would survive much file access from userland,
> BW> at least it can handle gzip-9 compression - well slow or course...
> BW> 
> BW> [55]momax# zpool list
> BW> NAME                    SIZE    USED   AVAIL    CAP  HEALTH     ALTROOT
> BW> backup                   87G   74.8G   12.2G    85%  ONLINE     -
> BW> 
> BW> This very low configurtion makes it even possible to think about one
> BW> of the bigger Soekris with 256-512M RAM, although I would suggest to
> BW> use 512M for a real use ZFS, but backup with an 4801 seems to be
> BW> possible.
> 
> Would you please post zfs-related settings you use? Thanks in advance.

Oh - it was just vm.kmem_size="134217728" and not _many things_.
If kmem uses that much RAM there is not much left for applications, but
it runs stable for doing the backups.
I noticed once that a zpool scrub requires additional kmem to normal
use, which might still overflow this box - it is hard on the limit
after all.

-- 
B.Walter                http://www.bwct.de      http://www.fizon.de
bernd_at_bwct.de           info_at_bwct.de            support_at_fizon.de
Received on Sat Jun 23 2007 - 17:14:28 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:13 UTC