Date: Thu, 11 Oct 2012 10:41:21 -0700 From: Adrian Chadd <adrian@freebsd.org> To: Adam McDougall <mcdouga9@egr.msu.edu> Cc: freebsd-current@freebsd.org Subject: Re: Buying recommendation for silent router/fileserver Message-ID: <CAJ-Vmo=LVNSZjOPzumZX0ZKrNkNANpCiyzNO2=p07ri7qk4tLA@mail.gmail.com> In-Reply-To: <5076F955.8070207@egr.msu.edu> References: <20121011145453.GU69724@acme.spoerlein.net> <20121011160521.GB40357@in-addr.com> <5076F955.8070207@egr.msu.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Did you ever file a PR for the slow SATA behaviour? Adrian On 11 October 2012 09:52, Adam McDougall <mcdouga9@egr.msu.edu> wrote: > On 10/11/12 12:05, Gary Palmer wrote: >> >> On Thu, Oct 11, 2012 at 04:54:53PM +0200, Ulrich Sp??rlein wrote: >>> >>> Hey guys, >>> >>> I need to replace an aging Pentium IV system that has been serving as my >>> router, access point, file- and mediaserver for quite some time now. The >>> replacement should have: >>> >>> - amd64 CPU (for ZFS, obviously) >>> - 2x GigE (igress, egress interfaces) >>> - some form of wlan interface (I currently use an Atheros based PCI card) >>> - eSATA for attaching a backup disk where I stream ZFS snapshots to >>> - serial port is always nice, for when I mess up an upgrade >>> - fan-less if possible >>> >>> So far, this here seems to fit the bill perfectly >>> http://www.fit-pc.com/web/fit-pc/intensepc/ >>> but pricing seems to defy any reality. >>> >>> It does not state directly which chipsets are used for Wifi and >>> Ethernet, the block diagram claims Ethernet chips to be Intel 82579 and >>> RTL8111D, but I don't trust that fully. >>> >>> For Wifi I can always fall back to sticking in a supported USB stick, >>> although that's kinda hacky. >>> >>> So how well is networking going to be supported by FreeBSD? Should I >>> just bite the bullet and find out? >> >> >> I'd recommend the Soekris net6501, but it's even more expensive than the >> intensepc (I suspect due to low hardware volumes but thats just a guess) >> >> http://soekris.com/products/net6501.html >> >> You also don't specify what kind of storage you need, which is obviously >> an important factor for a file/media server. >> >> Gary >> _______________________________________________ >> freebsd-current@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-current >> To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org" >> > > Be wary of the Soekris net6501, I bought three of the 1.6Ghz net6501-70 > model which has an Atom E-680 cpu (E series) and it compiles more than twice > as slow as a 1.6Ghz Atom N270 in an older netbook. Someone else running > Linux reported similar CPU slowness. As far as practical network > throughput, I could only get 100Mbit/sec with a simple HTTP download of a > file full of zeros, and OpenVPN could only push about 25Mbit/sec. As a > practical example of the CPU slowness, it takes about 1.5 minutes to compile > pkg on the N270 netbook and 5 minutes on the 6501 (around 4.5 if I use -j2). > A kernel compile took an hour. Unfortunately I had no idea this CPU > (possibly implementation?) was so slow before I purchased it, and I could > scarcely find evidence of it on google after hours of searching when I had > already discovered the issue. I was hoping to find some comparative > benchmarks between various Atom series but manufacturers generally don't do > that. > > Additionally, the total AHCI SATA write speed on the net6501 (in BSD only?) > has a strange 20MB/sec limitation but reads can go over 100MB/sec. If I > write to one disk I get 20MB/sec, if I write to both SATA disks I get > 10MB/sec each. Write is equally slow on a SSD. Both someone running > OpenBSD and I running FreeBSD reported the same symptoms to the soekris-tech > mailing list and received no useful replies towards getting that problem > solved. I tested the write speed briefly with Linux and it did not appear > to have the 20MB/sec limitation. I did confirm it was using MSI(-X?) with > boot -v. I think this hardware would need to fall into Alexander Motin's > hands to get anywhere with debugging the SATA speed issue. Since it seems > fine in Linux, maybe some day it can be fixed in BSD but I have no clue how > that limitation could happen. The disks I tested with are fine in normal > computers. > > _______________________________________________ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmo=LVNSZjOPzumZX0ZKrNkNANpCiyzNO2=p07ri7qk4tLA>