Re: [zfs-discuss] ZFS Project Hardware

2008-06-03 Thread Volker A. Brandt
 Timely discussion. I too am trying to build a stable yet inexpensive storage
 server for my home lab
[...]

 Other options are that I build a whitebox or buy a new PowerEdge or Sun
 X2200 etc

If this is really just a lab storage server then an X2100M2 will be
enough.  Just get the minimum spec, buy two 3.5 SATA-II disks
(I guess the sweet spot is 750GB right now), and buy 8GB of third
party memory to max out the box for ZFS.

Then set up a ZFS-rooted Nevada and you're in business.  Depending
on your requirements, you have slightly over 1.3TB capacity, or
about 690GB mirrored.

I have just such a machine and am very happy.  I do run S10U5
on it since I need the box for other things, too.  So I don't have
ZFS root.


HTH -- Volker
-- 

Volker A. Brandt  Consulting and Support for Sun Solaris
Brandt  Brandt Computer GmbH   WWW: http://www.bb-c.de/
Am Wiesenpfad 6, 53340 Meckenheim Email: [EMAIL PROTECTED]
Handelsregister: Amtsgericht Bonn, HRB 10513  Schuhgröße: 45
Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-06-02 Thread Erik Trimble
Keith Bierman wrote:
 On May 30, 2008, at 6:59 PM, Erik Trimble wrote:

   
 The only drawback of the older Socket 940 Opterons is that they don't
 support the hardware VT extensions, so running a Windows guest  
 under xVM
 on them isn't currently possible.
 



  From the VirtualBox manual, page 11

 • No hardware virtualization required. VirtualBox does not require  
 processor
 features built into newer hardware like VT-x (on Intel processors) or  
 AMD-V
 (on AMD processors). As opposed to many other virtualization  
 solutions, you
 can therefore use VirtualBox even on older hardware where these  
 features are
 not present. In fact, VirtualBox’s sophisticated software techniques  
 are typically
 faster than hardware virtualization, although it is still possible to  
 enable hard-
 ware virtualization on a per-VM basis. Only for some exotic guest  
 operating
 systems like OS/2, hardware virtualization is required.


 

 I've been running windows under OpenSolaris on an aged 32-bit Dell.  
 I'm morally certain it lacks the hardware support, and in any event,  
 the VBOX configuration is set to avoid using the VT extensions anyway.

 Runs fine. Not the fastest box on the planet ... but it's got limited  
 DRAM.

   

That is correct. VirtualBox does _not_ require the VT extensions.  I was 
referring to xVM, which I'm still taking as synonymous with the 
Xen-based system.  xVM _does_ require the VT hardware extensions to run 
guest OSes in an unmodified form, which currently includes all flavors 
of Windows.


-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-06-02 Thread Keith Bierman

On Jun 2, 2008, at 3:24 AM   6/2/, Erik Trimble wrote:

 Keith Bierman wrote:
 On May 30, 2008, at 6:59 PM, Erik Trimble wrote:


 The only drawback of the older Socket 940 Opterons is that they  
 don't
 support the hardware VT extensions, so running a Windows guest   
 under xVM
 on them isn't currently possible.



 That is correct. VirtualBox does _not_ require the VT extensions.   
 I was referring to xVM, which I'm still taking as synonymous with  
 the Xen-based system.  xVM _does_ require the VT hardware  
 extensions to run guest OSes in an unmodified form, which currently  
 includes all flavors of Windows.



Ah, Marketing rebranding befuddles again.

It's Sun xVM VirtualBox (tm) as best I can tell from sun.com. So I  
assumed you were using the xVM in generic sense, not as Xen vs.  
Virtual Box.


-- 
Keith H. Bierman   [EMAIL PROTECTED]  | AIM kbiermank
5430 Nassau Circle East  |
Cherry Hills Village, CO 80113   | 303-997-2749
speaking for myself* Copyright 2008




___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-06-02 Thread Tom Buskey
 (2) You want a 64-bit CPU.  So that probably rules
 out your P4 machines, 
 unless they were extremely late-model P4s with the
 EM64T features.   
 Given that file-serving alone is relatively low-CPU,
 you can get away 
 with practically any 64-bit capable CPU made in the
 last 4 years.

Assuming you're topped out at 4GB RAM by your motherboard, how much difference 
does 32bit vs 64bit make?

Some of the old (cheap) SCSI cards I have only have 32 bit drivers on x86.  The 
cards supported in the 64bit kernel I have to buy, maybe for more then I paid 
for my motherboard+cpu+RAM.

So I'm curious about that point.

 (5) external cases/enclosures are expensive, but
 nice. The bang-for-buck 
 is in the workgroup server case, which (besides
 being a PC case) 
 generally holds 8-10 drives for about $300 or so.

Even more bang for the buck:
   PC case with power supply
   42 SATA cables
   SATA to slot bracket cables
   disk drive power adapters to SATA power adapters
   1-2 120mm fans


Bolt the drives into the case.
Plug in the power adapters
Bolt the SATA to slot brackets in
Plug the SATA connectors into the drives.
Mount the fan(s) in the case so it pulls air across the drives.

Bolt the box up.  Now you have drives in a box with SATA ports out the back.

Now plug the 42 cables into your server in the SATA ports and feed them out a 
hole in the case.  I label the port numbers on the cables.

Plug the 42 cables into the SATA ports of the drives in a box.

Power the drives in a box up before your server.

I've found this works.  SATA isn't sensitive to setup as SCSI was in the past.
You did mention this was for a home lab right?  I've been running this way in 
my servers for 4 years.


 If the solution you really want is an external disk
 enclosure hooked to 
 some sort of a driver/head machine, check out
 used/off-lease IBM or HP 
 opteron workstations, which tend to go for $500 or
 so, loaded.  Sun v20z 
 and IBM e326m  1U rackmount servers are in the same
 price range.
 
 -- 
 Erik Trimble
 Java System Support
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-06-01 Thread Keith Bierman

On May 30, 2008, at 6:59 PM, Erik Trimble wrote:

 The only drawback of the older Socket 940 Opterons is that they don't
 support the hardware VT extensions, so running a Windows guest  
 under xVM
 on them isn't currently possible.



 From the VirtualBox manual, page 11

• No hardware virtualization required. VirtualBox does not require  
processor
features built into newer hardware like VT-x (on Intel processors) or  
AMD-V
(on AMD processors). As opposed to many other virtualization  
solutions, you
can therefore use VirtualBox even on older hardware where these  
features are
not present. In fact, VirtualBox’s sophisticated software techniques  
are typically
faster than hardware virtualization, although it is still possible to  
enable hard-
ware virtualization on a per-VM basis. Only for some exotic guest  
operating
systems like OS/2, hardware virtualization is required.




I've been running windows under OpenSolaris on an aged 32-bit Dell.  
I'm morally certain it lacks the hardware support, and in any event,  
the VBOX configuration is set to avoid using the VT extensions anyway.

Runs fine. Not the fastest box on the planet ... but it's got limited  
DRAM.



-- 
Keith H. Bierman   [EMAIL PROTECTED]  | AIM kbiermank
5430 Nassau Circle East  |
Cherry Hills Village, CO 80113   | 303-997-2749
speaking for myself* Copyright 2008




___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Marc Bevand
Marc Bevand m.bevand at gmail.com writes:
 
 What I hate about mobos with no onboard video is that these days it is 
 impossible to find cheap fanless video cards. So usually I just go headless.

Didn't finish my sentence: ...fanless and *power-efficient*.
Most cards consume 20+W when idle. This alone is a half or a
third of the idle power consumption of a small NAS.

-marc

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread SS
Timely discussion. I too am trying to build a stable yet inexpensive storage 
server for my home lab mostly for playing in the VM world as well as general 
data storage. I've considered several options ranging from the simple linux 
based NAS appliances to older EMC SANs. I finally decided to build an 
NFS/CIFS/iSCSI/(even FC target?) box going the opensolaris route with ZFS. What 
I'm trying to decide on is the appropriate hardware to build the storage 
server. I have:

- A couple of Dell Pentium 4 boxes
- A couple of old Ultra SPARC (ultra80 and ultra 10)
- D1000 array (but alas with old 36G drives)

Other options are that I build a whitebox or buy a new PowerEdge or Sun X2200 
etc use some kind of DAS such as Dell MD1000 (?) and use this box as the one 
and the only system (i.e. storage for PCs and my VM host). Of course this will 
be an expensive option.

Any recommendations on a decent setup for my purposes as well as a good SATA 
DAS? I haven't build a PC for at least 4 years so I'm not up to date on the 
processors, mobos, controller cards etc.

PS. Question for the gentleman who bought the external SATA disk array...how 
are you planning to connect it to the server?
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Erik Trimble
SS wrote:
 Timely discussion. I too am trying to build a stable yet inexpensive storage 
 server for my home lab mostly for playing in the VM world as well as general 
 data storage. I've considered several options ranging from the simple linux 
 based NAS appliances to older EMC SANs. I finally decided to build an 
 NFS/CIFS/iSCSI/(even FC target?) box going the opensolaris route with ZFS. 
 What I'm trying to decide on is the appropriate hardware to build the storage 
 server. I have:

 - A couple of Dell Pentium 4 boxes
 - A couple of old Ultra SPARC (ultra80 and ultra 10)
 - D1000 array (but alas with old 36G drives)

 Other options are that I build a whitebox or buy a new PowerEdge or Sun X2200 
 etc use some kind of DAS such as Dell MD1000 (?) and use this box as the one 
 and the only system (i.e. storage for PCs and my VM host). Of course this 
 will be an expensive option.

 Any recommendations on a decent setup for my purposes as well as a good SATA 
 DAS? I haven't build a PC for at least 4 years so I'm not up to date on the 
 processors, mobos, controller cards etc.

 PS. Question for the gentleman who bought the external SATA disk array...how 
 are you planning to connect it to the server

Well, a couple of things:

(1) we need to know more about your expected performance and use 
requirements before making a real recommendation

(2) You want a 64-bit CPU.  So that probably rules out your P4 machines, 
unless they were extremely late-model P4s with the EM64T features.   
Given that file-serving alone is relatively low-CPU, you can get away 
with practically any 64-bit capable CPU made in the last 4 years.

(3) As much as I love them, the ultra80  ultra10 are boat anchors now.  
Way too slow, way too power hungry, and not really useful.

(4)  High capacity on a budget means SATA drives.  You can use small 
SCSI drives for certain performance-sensitive applications and not get 
creamed in the pocketbook (the D1000 is kinda interesting for this), but 
you need some form of SATA to get the big GB/$ benefits.

(5) external cases/enclosures are expensive, but nice. The bang-for-buck 
is in the workgroup server case, which (besides being a PC case) 
generally holds 8-10 drives for about $300 or so.

There's lots of not-quite-optimal-but-still-really-good solutions out 
there on the used/recycled market, so if you don't need something 
perfect (or a warranty), the price is really nice.


If the solution you really want is an external disk enclosure hooked to 
some sort of a driver/head machine, check out used/off-lease IBM or HP 
opteron workstations, which tend to go for $500 or so, loaded.  Sun v20z 
and IBM e326m  1U rackmount servers are in the same price range.

-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Justin Vassallo
Bob said:
 SATA enterprise drives seem more like a gimmick than anything else. 
 Perhaps the warranty is longer and they include a tiny bit more smarts 
 in the firmware.

WD supply enterprise class SATA drives whose prevailing feature is a low
TLER (RE series). This makes the drive report a failed block quickly, rather
than trying to recover the blocks for minutes. In a consumer PC, the drive's
heroic attempts can save the day. In a raid setup, however, this means that
a drive will lock a raided fs until it times out. I would rather have the
drive report as failed and let the raid take care of recovering the data.

Even if there were no further technical reasons, this feature alone is a
great benefit for using these SATA drives in the enterprise

justin


smime.p7s
Description: S/MIME cryptographic signature
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Brandon High
On Sat, May 31, 2008 at 7:06 AM, Justin Vassallo
[EMAIL PROTECTED] wrote:
 WD supply enterprise class SATA drives whose prevailing feature is a low
 TLER (RE series). This makes the drive report a failed block quickly, rather
 than trying to recover the blocks for minutes. In a consumer PC, the drive's

The same feature can be enabled on WD's consumer SATA drives. Google
for wdtler.zip.

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Brandon High
On Fri, May 30, 2008 at 10:58 PM, Marc Bevand [EMAIL PROTECTED] wrote:
 For example the Abit AB9 Pro (about $80-90) comes with 10 SATA ports (9
 internal + 1 internal): 6 from the ICH8R chipset (driver: ahci), 2 from a
 JMB363 chip (driver: ahci in snv_82 and above, see bug 6645543), and 2 from a
 SiI3132 chip (driver: si3124).

I had hoped to get a system with on board ports, but hadn't found one
with more than 6. Thanks for the pointer!

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-31 Thread Richard Elling
Brandon High wrote:
 On Fri, May 30, 2008 at 10:58 PM, Marc Bevand [EMAIL PROTECTED] wrote:
   
 For example the Abit AB9 Pro (about $80-90) comes with 10 SATA ports (9
 internal + 1 internal): 6 from the ICH8R chipset (driver: ahci), 2 from a
 JMB363 chip (driver: ahci in snv_82 and above, see bug 6645543), and 2 from a
 SiI3132 chip (driver: si3124).
 

 I had hoped to get a system with on board ports, but hadn't found one
 with more than 6. Thanks for the pointer!

   

Look for motherboards using the NVidia 680[ai].
http://en.wikipedia.org/wiki/NForce_600
 -- richard

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Orvar Korvar
Im using the AOC card with 8 SATA-2 ports too. It got detected automatically 
during Solaris install. Works great. And it is cheap. Ive heard that it is the 
same chipset as used in X4500 thumper with 48 drives? 

In a PCI, the PCI bottle necks at ~150MB/sec, or so.

In a PCI-X slot, you will reach something like 1.5GB/sec which should suffice 
for most needs. Maybe it is cheaper to buy that card + PCI-X motherboard (only 
found on server mobos) than buying a SAS or PCI-express, if you want to achieve 
speed?
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Brandon High
On Fri, May 30, 2008 at 12:48 PM, Orvar Korvar
[EMAIL PROTECTED] wrote:
 In a PCI-X slot, you will reach something like 1.5GB/sec which should suffice 
 for most needs. Maybe it is cheaper to buy that card + PCI-X motherboard 
 (only found on server mobos) than buying a SAS or PCI-express, if you want to 
 achieve speed?

That's my thought as well. I'm going to be putting together a home NAS
based on OpenSolaris using the following:
1   SUPERMICRO CSE-743T-645B Black Chassis  
1   ASUS M2N-LR AM2 NVIDIA nForce Professional 3600 ATX Server Motherboard  
1   SUPERMICRO AOC-SAT2-MV8 64-bit PCI-X133MHz SATA Controller Card 
1   AMD Athlon X2 4850e 2.5GHz Socket AM2 45W Dual-Core Processor Model
ADH4850DOBOX
1   Crucial 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 667 (PC2 5300) ECC
Unbuffered Dual Channel Kit Server Memory Model CT2KIT25672AA667
8   Western Digital Caviar GP WD10EACS 1TB 5400 to 7200 RPM SATA
3.0Gb/s Hard Drive

Subtotal:   $2,386.88

I may get another drive for the OS as well, or boot off of a
CF-card/IDE adapter like this one:
http://www.newegg.com/Product/Product.aspx?Item=N82E16812186038

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Erik Trimble
Brandon High wrote:
 On Fri, May 30, 2008 at 12:48 PM, Orvar Korvar
 [EMAIL PROTECTED] wrote:
   
 In a PCI-X slot, you will reach something like 1.5GB/sec which should 
 suffice for most needs. Maybe it is cheaper to buy that card + PCI-X 
 motherboard (only found on server mobos) than buying a SAS or PCI-express, 
 if you want to achieve speed?
 

 That's my thought as well. I'm going to be putting together a home NAS
 based on OpenSolaris using the following:
 1 SUPERMICRO CSE-743T-645B Black Chassis  
 1 ASUS M2N-LR AM2 NVIDIA nForce Professional 3600 ATX Server Motherboard  
 1 SUPERMICRO AOC-SAT2-MV8 64-bit PCI-X133MHz SATA Controller Card 
 1 AMD Athlon X2 4850e 2.5GHz Socket AM2 45W Dual-Core Processor Model
 ADH4850DOBOX
 1 Crucial 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 667 (PC2 5300) ECC
 Unbuffered Dual Channel Kit Server Memory Model CT2KIT25672AA667
 8 Western Digital Caviar GP WD10EACS 1TB 5400 to 7200 RPM SATA
 3.0Gb/s Hard Drive

 Subtotal: $2,386.88

 I may get another drive for the OS as well, or boot off of a
 CF-card/IDE adapter like this one:
 http://www.newegg.com/Product/Product.aspx?Item=N82E16812186038

 -B
   

One thought on this:for a small server, which is unlikely to ever be 
CPU bound, I would suggest looking for an older dual-Socket 940 Opteron 
motherboard.  They almost all have many PCI-X slots, and single-core 
Opterons are dirt cheap.  PC3200 DDR1 ECC ram is also cheap too.

Tyan/Supermicro dual-Socket 940 motherboard:   $200
Opteron 252 (2.4Ghz) + heatsink   $75
Opteron 280 (dual-core 2.4Ghz) + heatsink  $180
4x1GB ECC DDR1 PC3200 ram $150


The only drawback of the older Socket 940 Opterons is that they don't 
support the hardware VT extensions, so running a Windows guest under xVM 
on them isn't currently possible.

-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Brandon High
On Fri, May 30, 2008 at 5:59 PM, Erik Trimble [EMAIL PROTECTED] wrote:
 One thought on this:for a small server, which is unlikely to ever be CPU
 bound, I would suggest looking for an older dual-Socket 940 Opteron
 motherboard.  They almost all have many PCI-X slots, and single-core
 Opterons are dirt cheap.  PC3200 DDR1 ECC ram is also cheap too.

 Tyan/Supermicro dual-Socket 940 motherboard:   $200
 Opteron 252 (2.4Ghz) + heatsink   $75
 Opteron 280 (dual-core 2.4Ghz) + heatsink  $180
 4x1GB ECC DDR1 PC3200 ram $150

I've actually done the math, and while your statement may have been
true a few months ago, it doesn't appear to be the case anymore. Right
now Newegg shows the Opteron 285 for $389, or the 880 for $694.

The cheapest ECC DDR RAM is $40 per GB.

The least expensive Socket 940 board with a PCI-X slot is the TYAN
S2881UG2NR at $419.

Call it $960 (with a single 285 cpu) vs. $399 for the AM2 pieces.

I'd check prices on a single socket 939 Opteron with a suitable
motherboard, but neither appear to be available anymore.

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Tim
USED hardware is your friend :)  He wasn't quoting new prices.



On Fri, May 30, 2008 at 8:54 PM, Brandon High [EMAIL PROTECTED] wrote:

 On Fri, May 30, 2008 at 5:59 PM, Erik Trimble [EMAIL PROTECTED]
 wrote:
  One thought on this:for a small server, which is unlikely to ever be
 CPU
  bound, I would suggest looking for an older dual-Socket 940 Opteron
  motherboard.  They almost all have many PCI-X slots, and single-core
  Opterons are dirt cheap.  PC3200 DDR1 ECC ram is also cheap too.
 
  Tyan/Supermicro dual-Socket 940 motherboard:   $200
  Opteron 252 (2.4Ghz) + heatsink   $75
  Opteron 280 (dual-core 2.4Ghz) + heatsink  $180
  4x1GB ECC DDR1 PC3200 ram $150

 I've actually done the math, and while your statement may have been
 true a few months ago, it doesn't appear to be the case anymore. Right
 now Newegg shows the Opteron 285 for $389, or the 880 for $694.

 The cheapest ECC DDR RAM is $40 per GB.

 The least expensive Socket 940 board with a PCI-X slot is the TYAN
 S2881UG2NR at $419.

 Call it $960 (with a single 285 cpu) vs. $399 for the AM2 pieces.

 I'd check prices on a single socket 939 Opteron with a suitable
 motherboard, but neither appear to be available anymore.

 -B

 --
 Brandon High [EMAIL PROTECTED]
 The good is the enemy of the best. - Nietzsche
 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Brandon High
On Fri, May 30, 2008 at 6:57 PM, Tim [EMAIL PROTECTED] wrote:
 USED hardware is your friend :)  He wasn't quoting new prices.

Not really an apples-to-apples comparison then, is it? Cruising eBay
for parts isn't my idea of reproducible or supportable.

Sure, an older server could possibly fall into my car's trunk as I
leave work one day, but that's not something I'd consider either.

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Erik Trimble
Brandon High wrote:
 On Fri, May 30, 2008 at 5:59 PM, Erik Trimble [EMAIL PROTECTED] wrote:
   
 One thought on this:for a small server, which is unlikely to ever be CPU
 bound, I would suggest looking for an older dual-Socket 940 Opteron
 motherboard.  They almost all have many PCI-X slots, and single-core
 Opterons are dirt cheap.  PC3200 DDR1 ECC ram is also cheap too.

 Tyan/Supermicro dual-Socket 940 motherboard:   $200
 Opteron 252 (2.4Ghz) + heatsink   $75
 Opteron 280 (dual-core 2.4Ghz) + heatsink  $180
 4x1GB ECC DDR1 PC3200 ram $150
 

 I've actually done the math, and while your statement may have been
 true a few months ago, it doesn't appear to be the case anymore. Right
 now Newegg shows the Opteron 285 for $389, or the 880 for $694.

 The cheapest ECC DDR RAM is $40 per GB.

 The least expensive Socket 940 board with a PCI-X slot is the TYAN
 S2881UG2NR at $419.

 Call it $960 (with a single 285 cpu) vs. $399 for the AM2 pieces.

 I'd check prices on a single socket 939 Opteron with a suitable
 motherboard, but neither appear to be available anymore.

 -B
   
Prices I quoted are for used, or excess stock from eBay.  There's still 
a lot of reman stuff out there which is new, and very cheap.
http://www.pricewatch.com/  is also a good place to look for 
not-quite-retail-packaged.


A quick eBay jaunt yields (buy-it-now prices including shipping):

$70   Opteron 270 (2.0Ghz) - currently the sweet spot for dual-core opterons
$20   Opteron 252 (2.6ghz) - the sweet spot for single-core opterons
$30   Opteron heatsink  (you can probably do better at a local computer 
store)
$160  4-pack 1GB Viking-branded PC3200 ECC DDR1 Dimms
$200   Tyan Thunder K8WE motherboard
$120   Supermicro H8DAE-B motherboard

As an aside, there's someone selling a Tyan Thunder K8S with two opteron 
246 (1.8Ghz) and heat sinks, all for $150.


Also, getting an Opteron motherboard tends to get you better extras, 
like more gigabit ethernet ports, built-in VGA, and either SCSI or lots 
of SATA ports.  Not to mention usually at least twice as many PCI-X slots.

But, your mileage may vary. It was just a suggestion.



-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Erik Trimble
Brandon High wrote:
 On Fri, May 30, 2008 at 6:57 PM, Tim [EMAIL PROTECTED] wrote:
   
 USED hardware is your friend :)  He wasn't quoting new prices.
 

 Not really an apples-to-apples comparison then, is it? Cruising eBay
 for parts isn't my idea of reproducible or supportable.

 Sure, an older server could possibly fall into my car's trunk as I
 leave work one day, but that's not something I'd consider either.

 -B

   
That said, much of what is available on eBay and from computer 
recyclers really isn't USED.  It's surplus inventory, overstock, etc.  
Much of it still in the OEM or retail box.

Heck, I know a bunch of folks selling BRAND_NEW  IBM e326m machines, 
complete with 1-year IBM factory warranty, still-in-the-box, for $400   
(1 Opteron 280/1GB RAM/1U rackmount chassis).   It takes a bit of 
hunting, but you'll find that rummaging around the Internet can still 
get you new parts for obsolete machines, at cut-rate pricing.

-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-30 Thread Marc Bevand
Brandon High bhigh at freaks.com writes:
 
 I'm going to be putting together a home NAS
 based on OpenSolaris using the following:
 1 SUPERMICRO CSE-743T-645B Black Chassis  
 1 ASUS M2N-LR AM2 NVIDIA nForce Professional 3600 ATX Server Motherboard  
 1 SUPERMICRO AOC-SAT2-MV8 64-bit PCI-X133MHz SATA Controller Card 
 1 AMD Athlon X2 4850e 2.5GHz Socket AM2 45W Dual-Core Processor Model
 ADH4850DOBOX
 1 Crucial 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 667 (PC2 5300) ECC
 Unbuffered Dual Channel Kit Server Memory Model CT2KIT25672AA667
 8 Western Digital Caviar GP WD10EACS 1TB 5400 to 7200 RPM SATA
 3.0Gb/s Hard Drive
 
 Subtotal: $2,386.88

You could get a $200 cheaper, more power-efficient, and more performant config 
by buying a high SATA port-count desktop-class mobo instead of a server one + 
AOC-SAT2-MV8. 

For example the Abit AB9 Pro (about $80-90) comes with 10 SATA ports (9 
internal + 1 internal): 6 from the ICH8R chipset (driver: ahci), 2 from a 
JMB363 chip (driver: ahci in snv_82 and above, see bug 6645543), and 2 from a 
SiI3132 chip (driver: si3124).

All these drivers should be rock-solid. Performance-wise you should be able to 
max out your 8 disks' max read/write throughput at the same time (but see 
http://opensolaris.org/jive/thread.jspa?threadID=54481 there is usually a 
bottleneck of 150 MB/s per PCI-E lane, this apply to the JMB363 and SiI3132).

Downside: loss of upgradability by having onboard SATA controllers. No onboard 
video. And it's an Intel mobo. Intel's prices for low-power processors (below 
~50W) are higher than AMD's, especially for dual-core ones. But something only 
slightly more power-hungry than your 45W AMD is the Pentium E2220 (2.4GHz 
dual-core 65W). Most likely your NAS will spend 90+% of its time idle so there 
wouldn't be a constant 20W power diff between the 2 configs.

What I hate about mobos with no onboard video is that these days it is 
impossible to find cheap fanless video cards. So usually I just go headless.

-marc

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-29 Thread Brian Hechinger
On Wed, May 28, 2008 at 12:01:36PM -0400, Bill McGonigle wrote:
 On May 28, 2008, at 05:11, James Andrewartha wrote:
 
 That's not a huge price difference when building a server - thanks  
 for the pointer.  Are there any 'gotchas' the list can offer when  
 using a SAS card with SATA drives?   I've been told that SATA drives  
 can have a lower MTBF than SAS drives (by a guy working QA for  
 BigDriveCo), but ZFS helps keep the I in RAID.

I'm running 3 (used to be 4, but I repurposed that drive) 500GB Seagate
SATA disks on an LSI SAS3080X in a RAIDZ1 pool in my Ultra80 and it's
been working great.  The only 'gothca' that I can think of is the loss
of the ability to run more than one drive per channel, but I guess I can
live with that. :)

I got my SAS3080X for, uhm, let's see, including shipping and the SAS to
4 cable SATA breakout cable, it was less than $100 off of ebay, probably
closer to $80.

I don't know prices on the PCIe version of those cards on ebay though.
Probably more expensive as everyone wants PCIe these days.

-brian
-- 
Coding in C is like sending a 3 year old to do groceries. You gotta
tell them exactly what you want or you'll end up with a cupboard full of
pop tarts and pancake mix. -- IRC User (http://www.bash.org/?841435)
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-29 Thread Mathew P
I've had a RAIDZ/ZFS File Server since Update 2, so I thought I'd share my 
setup.

Opteron FX-51 (2.3Ghz, Socket 939)
Asus SK8N
4x 512MB EBB Unbuffered DDR1 Memory
2x Skymaster PCI-X 4 Port SATA (based on SI3114 Chipset).  currently deployed 
over 2x PCI ports on the motherboard.
1x Intel 10/100 NIC PCI.
8x 320GB Western Digital SATA Drives

As you can see, I'm sharing the PCI bus for both controllers and my NIC.  So 
the speed isn't very fast (10MB/s) from a Windows XP client through Samba.  

I'm considering changing motherboard to Asus K8N-LR, which will allow me to use 
both PCI-X slots, and a dedicated Intel Gigabit NIC in the PCIe slot.  Which 
will dramatically speed things up.
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread James Andrewartha
Erik Trimble wrote:
 On a related note - does anyone know of a good Solaris-supported 4+ port 
 SATA card for PCI-Express?  Preferably 1x or 4x slots...

 From what I can tell, all the vendors are only making SAS controllers for 
PCIe with more than 4 ports. Since SAS supports SATA, I guess they don't see 
much point in doing SATA-only controllers.

For example, the LSI SAS3081E-R is $260 for 8 SAS ports on 8x PCIe, which is 
somewhat more expensive than the almost equivalent PCI-X LSI SAS3080X-R 
which is as low as $180.

For those downthread looking for full RAID controllers with battery backup 
RAM, Areca (who formerly specialised in SATA controlers) now do SAS RAID at 
reasonable prices, and have Solaris drivers.

-- 
James Andrewartha
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Bill McGonigle
On May 28, 2008, at 05:11, James Andrewartha wrote:

  From what I can tell, all the vendors are only making SAS  
 controllers for
 PCIe with more than 4 ports. Since SAS supports SATA, I guess they  
 don't see
 much point in doing SATA-only controllers.

 For example, the LSI SAS3081E-R is $260 for 8 SAS ports on 8x PCIe,  
 which is
 somewhat more expensive than the almost equivalent PCI-X LSI  
 SAS3080X-R
 which is as low as $180.

That's not a huge price difference when building a server - thanks  
for the pointer.  Are there any 'gotchas' the list can offer when  
using a SAS card with SATA drives?   I've been told that SATA drives  
can have a lower MTBF than SAS drives (by a guy working QA for  
BigDriveCo), but ZFS helps keep the I in RAID.

 For those downthread looking for full RAID controllers with battery  
 backup
 RAM, Areca (who formerly specialised in SATA controlers) now do SAS  
 RAID at
 reasonable prices, and have Solaris drivers.

I've seen posts about misery with the sil and marvell drivers from  
about a year ago; is there a good way to pound an opensolaris driver  
to find its holes, in a ZFS context?  On one hand I'd guess it  
shouldn't be too hard to simulate different kinds of loads, but on  
the other hand, if that were easy, the drivers' authors would have  
done that before unleashing buggy code on the masses.

Thanks,
-Bill

-
Bill McGonigle, Owner   Work: 603.448.4440
BFC Computing, LLC  Home: 603.448.1668
[EMAIL PROTECTED]   Cell: 603.252.2606
http://www.bfccomputing.com/Page: 603.442.1833
Blog: http://blog.bfccomputing.com/
VCard: http://bfccomputing.com/vcard/bill.vcf

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Richard Elling
Bill McGonigle wrote:
 On May 28, 2008, at 05:11, James Andrewartha wrote:

   
  From what I can tell, all the vendors are only making SAS  
 controllers for
 PCIe with more than 4 ports. Since SAS supports SATA, I guess they  
 don't see
 much point in doing SATA-only controllers.

 For example, the LSI SAS3081E-R is $260 for 8 SAS ports on 8x PCIe,  
 which is
 somewhat more expensive than the almost equivalent PCI-X LSI  
 SAS3080X-R
 which is as low as $180.
 

 That's not a huge price difference when building a server - thanks  
 for the pointer.  Are there any 'gotchas' the list can offer when  
 using a SAS card with SATA drives?   I've been told that SATA drives  
 can have a lower MTBF than SAS drives (by a guy working QA for  
 BigDriveCo), but ZFS helps keep the I in RAID.
   

There are BigDriveCos which sell enterprise-class SATA drives.
Since the mechanics are the same, the difference is in the electronics
and software.  Vote with your pocketbook for the enterprise-class
products.
 -- richard

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Keith Bierman

On May 28, 2008, at 10:27 AM   5/28/, Richard Elling wrote:

 Since the mechanics are the same, the difference is in the electronics


In my very distant past, I did QA work for an electronic component  
manufacturer. Even parts which were identical were expected to  
behave quite differently ... based on population statistics. That is,  
the HighRel MilSpec parts were from batches with no failures (even  
under very harsh conditions beyond the normal operating mode, and all  
tests to destruction showed only the expected failure modes) and the  
hobbyist grade components were those whose cohort *failed* all the  
testing (and destructive testing could highlight abnormal failure  
modes).

I don't know that drive builders do the same thing, but I'd kinda  
expect it.

-- 
Keith H. Bierman   [EMAIL PROTECTED]  | AIM kbiermank
5430 Nassau Circle East  |
Cherry Hills Village, CO 80113   | 303-997-2749
speaking for myself* Copyright 2008




___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread James Andrewartha
On Wed, 2008-05-28 at 10:34 -0600, Keith Bierman wrote:
 On May 28, 2008, at 10:27 AM   5/28/, Richard Elling wrote:
 
  Since the mechanics are the same, the difference is in the electronics
 
 
 In my very distant past, I did QA work for an electronic component  
 manufacturer. Even parts which were identical were expected to  
 behave quite differently ... based on population statistics. That is,  
 the HighRel MilSpec parts were from batches with no failures (even  
 under very harsh conditions beyond the normal operating mode, and all  
 tests to destruction showed only the expected failure modes) and the  
 hobbyist grade components were those whose cohort *failed* all the  
 testing (and destructive testing could highlight abnormal failure  
 modes).
 
 I don't know that drive builders do the same thing, but I'd kinda  
 expect it.

Seagate's ES.2 has a higher MBTF than the equivalent consumer drive, so
you're probably right. Western Digital's RE2 series (which my work uses)
comes with a 5 year warranty, compared to 3 years for the consumer
versions. The RE2 also have firmware with Time-Limited Error Recovery,
which reports errors promptly, letting the higher-level RAID do data
recovery. Both have improved vibration tolerance through firmware
tweaks. And if you want 10krpm, I think WD's VelociRaptor counts.
http://www.techreport.com/articles.x/13732
http://www.techreport.com/articles.x/13253
http://www.techreport.com/articles.x/14583
http://www.storagereview.com/ is promising some SSD benchmarks soon.

James Andrewartha
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Brandon High
On Wed, May 28, 2008 at 9:27 AM, Richard Elling [EMAIL PROTECTED] wrote:
 There are BigDriveCos which sell enterprise-class SATA drives.
 Since the mechanics are the same, the difference is in the electronics
 and software.  Vote with your pocketbook for the enterprise-class
 products.

CMU released a study comparing the MTBF enterprise class drive with
consumer drives, and found no real differences.

From the study:
In our data sets, the replacement rates of SATA disks are not worse
than the replacement rates of SCSI or FC disks. This may indicate that
disk-independent factors, such as operating conditions, usage and
environmental factors affect replacement rates more than component
specific factors.

Google has also released a similar study on drive reliability.
Google's sample size is considerably larger than CMU's as well.
There's a blurb here: http://news.bbc.co.uk/2/hi/technology/6376021.stm
Full results here: http://research.google.com/archive/disk_failures.pdf

-B

-- 
Brandon High [EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Bob Friesenhahn
On Wed, 28 May 2008, Brandon High wrote:
 CMU released a study comparing the MTBF enterprise class drive with
 consumer drives, and found no real differences.

That should really not be a surprise.  Chips are chips and in the 
economies of scale, as few chips will be be used as possible.  The 
quality of manufacture could vary, but this is likely more dependent 
on the manufacturer than the product line.  Manufacturers who produce 
crummy products don't last very long.

True enterprise drives (SCSA, SAS, FC) have much lower media read 
error rates by an factor of 10 and more tolerance to vibration and 
temperature.  They also have much lower storage capacity and much 
better seek and I/O performance.  Failure to read a block is not a 
failure of the drive so this won't be considered by any study which 
only considers drive replacement.

SATA enterprise drives seem more like a gimmick than anything else. 
Perhaps the warranty is longer and they include a tiny bit more smarts 
in the firmware.

Bob
==
Bob Friesenhahn
[EMAIL PROTECTED], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,http://www.GraphicsMagick.org/

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-28 Thread Richard Elling
http://blogs.sun.com/relling/entry/adaptec_webinar_on_disks_and
 -- richard

Bob Friesenhahn wrote:
 On Wed, 28 May 2008, Brandon High wrote:
   
 CMU released a study comparing the MTBF enterprise class drive with
 consumer drives, and found no real differences.
 

 That should really not be a surprise.  Chips are chips and in the 
 economies of scale, as few chips will be be used as possible.  The 
 quality of manufacture could vary, but this is likely more dependent 
 on the manufacturer than the product line.  Manufacturers who produce 
 crummy products don't last very long.

 True enterprise drives (SCSA, SAS, FC) have much lower media read 
 error rates by an factor of 10 and more tolerance to vibration and 
 temperature.  They also have much lower storage capacity and much 
 better seek and I/O performance.  Failure to read a block is not a 
 failure of the drive so this won't be considered by any study which 
 only considers drive replacement.

 SATA enterprise drives seem more like a gimmick than anything else. 
 Perhaps the warranty is longer and they include a tiny bit more smarts 
 in the firmware.

   
 Bob
 ==
 Bob Friesenhahn
 [EMAIL PROTECTED], http://www.simplesystems.org/users/bfriesen/
 GraphicsMagick Maintainer,http://www.GraphicsMagick.org/

 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
   

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-27 Thread Bryan Wagoner
I'm using a gigabyte I-RAM card with cheap memory for my slog device with great 
results. Of course I don't have as much memory as you do in my project box.  I 
also want to use the left over space on the I-ram and dual purpose it for a 
readzilla cache device and slog.  Picked it up off ebay along with some 
computer guts and an nsc-314s 3u 14 hot swap drive rackmount case.

Like everyone else, I have been spending hours trying to find a supported high 
capacity SATA card that supports PCI-Express. I wish someone would make a 
driver for that adaptec card mentioned in this thread.  it is very reasonably 
priced for a project box. Everything else that fits that category and is 
supported seems to be $320 or much more.
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Marc Bevand
Tim tim at tcsac.net writes:
 
 So we're still stuck the same place we were a year ago.  No high port
 count pci-E compatible non-raid sata cards.  You'd think with all the
 demand SOMEONE would've stepped up to the plate by now.  Marvell, cmon ;)

Here is a 6-port SATA PCI-Express x1 controller for $70: [1]. I don't know who 
makes this card, but from the picture it is apparently based on a SiI3114 chip 
behind a PCI-E to PCI bridge. I also don't know how they get 6 ports total 
when this chip is known to only provide 4 ports.  Downsides: SATA 1.5 Gbps 
only; 4 of the ports are external (eSATA cables required); and don't expect to 
break throughput records because the bottleneck will be the internal PCI bus 
(33 MHz or 66 MHz: 133 or 266 MB/s theoretical hence 100 or 200 MB/s practical 
peak throughput shared between the 6 drives).

I also know Lycom, who is selling a 4-port PCI-E x8 card based on the Silicon 
Image SiI3124 chip and a PCI-E to PCI-X bridge [2]. I am unable to find a 
vendor for this card though. I heard about Lycom through the vendor list on 
sata-io.org.

Regarding Marvell, their website is completely useless as they provide almost 
no tech info regarding their SATA products, but according to a wikipedia 
article [3] they have three PCI-E to SATA 3.0 Gbps host controllers:

  o 88SE6141: 4-port (AHCI ?)
  o 88SE6145: 4-port (AHCI according to the Linux driver source code)
  o 88SX7042: 4-port (non-AHCI)

The 6141 and 6145 appear to be mostly used as onboard SATA controllers 
according to [3]. The 7042 can be found on some Adaptec and Highpoint cards 
according to [4], but they are probably expensive and come with this thing 
called hardware RAID that most of us don't need :)

Overall, like you I am frustrated by the lack of non-RAID inexpensive native 
PCI-E SATA controllers.

-marc

[1] http://cooldrives.com/ss42chesrapc.html
[2] http://www.lycom.com.tw/PE124R5.htm
[3] http://en.wikipedia.org/wiki/List_of_Marvell_Technology_Group_chipsets
[4] 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/ata/sata_mv.c;hb=HEAD


___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Kyle McDonald
Marc Bevand wrote:

 Overall, like you I am frustrated by the lack of non-RAID inexpensive native 
 PCI-E SATA controllers.


   
Why non-raid? Is it cost?

Personally I'm interested in a high port count RAID card, with as much 
battery-backed cache RAM as possible, and that can export as many LUNS 
as it can handle drives. I want a card like that so that I can give ZFS 
as many single drive RAID 0 luns that have battery-backed write caches 
as possible.

I know it won't be cheap, but it should perform really good.

In my current machines (IBM x346's), I'm stuck with U320 scsi internal 
for now, but I have the 256MB internal battery backed '7k' RAID card, 
and I've made 5 1 disk RAID0 LUNs to get the benefits of the write cache 
on the card but still let ZFS have the benefits of a many disk JBOD.

I'm on the look out for SATA RAID card with 1-4GB of battery-backed 
Cache to redo this config with 10-24 SATA drives.
I'll probably end up with multiple cards, since 1GB caches seem to be 
the most I've found.

   -Kyle
 [1] http://cooldrives.com/ss42chesrapc.html
 [2] http://www.lycom.com.tw/PE124R5.htm
 [3] http://en.wikipedia.org/wiki/List_of_Marvell_Technology_Group_chipsets
 [4] 
 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/ata/sata_mv.c;hb=HEAD


 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
   

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Marc Bevand
Kyle McDonald KMcDonald at Egenera.COM writes:
 Marc Bevand wrote:
 
  Overall, like you I am frustrated by the lack of non-RAID inexpensive
  native PCI-E SATA controllers.

 Why non-raid? Is it cost?

Primarily cost, reliability (less complex hw = less hw that can fail),
and serviceability (no need to rebuy the exact same raid card model
when it fails, any SATA controller will do).

If you want good write performance, instead of adding N GB of cache memory
to a disk controller, add N*5 or N*10 GB of system memory (DDR2 is maybe
1/5th or 1/10th cheaper per GB, and the OS already uses main memory to
cache disk writes).

-marc

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Bob Friesenhahn
On Sun, 25 May 2008, Marc Bevand wrote:

 Primarily cost, reliability (less complex hw = less hw that can fail),
 and serviceability (no need to rebuy the exact same raid card model
 when it fails, any SATA controller will do).

As long as the RAID is self-contained on the card, and the disks are 
exported as JBOD, then you should be able to replace the card with any 
adaptor supporting at least as many ports.

 If you want good write performance, instead of adding N GB of cache memory
 to a disk controller, add N*5 or N*10 GB of system memory (DDR2 is maybe
 1/5th or 1/10th cheaper per GB, and the OS already uses main memory to
 cache disk writes).

Something tells me that Kyle may know what he is talking about.  More 
system RAM does not help synchronous writes go much faster.  It does 
help with asynchronous writes, but only for intermittent or relatively 
slow write loads.  What makes the synchronous writes go faster is for 
the data to be queued as fast as possible to non-volatile media (e.g. 
NV write cache) so that the ZFS write operation can return right away. 
ZFS loads up the drives according to the current amount of I/O wait 
for the device.  If the device accepts data faster, then ZFS returns 
faster, and the client application (e.g. NFS or database) can run 
again.

Bob
==
Bob Friesenhahn
[EMAIL PROTECTED], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,http://www.GraphicsMagick.org/

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Rob
   More system RAM does not help synchronous writes go much faster.

agreed, but it does make sure all the asynchronous writes are
batched and the tgx group isn't committed early keeping everything
synchronous. (default batch is every 5 sec)

  If you want good write performance, instead of adding N GB of cache memory
  to a disk controller, add N*5 or N*10 GB of system memory (DDR2 is maybe

it kinda comes down to number of spindles in the end, even
if you have a huge log device, or huge system ZIL, it needs
to get to the disks (synchronous) in the end.

the rules don't change with zfs, the system with the most
vdevs wins :-)

Rob

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Kyle McDonald
Marc Bevand wrote:
 Kyle McDonald KMcDonald at Egenera.COM writes:
   
 Marc Bevand wrote:
 
 Overall, like you I am frustrated by the lack of non-RAID inexpensive
 native PCI-E SATA controllers.
   
 Why non-raid? Is it cost?
 

 Primarily cost, reliability (less complex hw = less hw that can fail),
 and serviceability (no need to rebuy the exact same raid card model
 when it fails, any SATA controller will do).

 If you want good write performance, instead of adding N GB of cache memory
 to a disk controller, add N*5 or N*10 GB of system memory (DDR2 is maybe
 1/5th or 1/10th cheaper per GB, and the OS already uses main memory to
 cache disk writes).

   
I've already maxed the machines out with 16GB. The RAID cache seemed the 
next step, and still cheaper than a SSD ZIL device. Though that I think 
would be the next step.

Since NFS is the primary way I intend to use this, the the battery 
backed RAM allows the sync requests to return much sooner than straight 
JBOD would.
At least that's my understanding.

  -Kyle

 -marc

 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
   

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread J. David Beutel
Which of these SATA controllers have people been able to use with SMART 
and ZFS boot in Solaris?

Cheers,
11011011
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-25 Thread Chris Siebenmann
|  Primarily cost, reliability (less complex hw = less hw that can
|  fail), and serviceability (no need to rebuy the exact same raid card
|  model when it fails, any SATA controller will do).
|
| As long as the RAID is self-contained on the card, and the disks are
| exported as JBOD, then you should be able to replace the card with any
| adaptor supporting at least as many ports.

 I believe it's common for PC-level hardware RAID cards to save the RAID
configuration on the disks themselves, which takes a bit of space and
(if it's done at the start of the disk) may make the disk unrecognizable
by standard tools, even with a JBOD setting.

 The vendors presumably do this, among other reasons, so that replacing
a dead controller doesn't require your operating system and so on to be
running in order to upload a saved configuration or the like.

- cks
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Aaron Blew
I've had great luck with my Supermicro AOC-SAT2-MV8 card so far.  I'm
using it in an old PCI slot, so it's probably not as fast as it could
be, but it worked great right out of the box.

-Aaron


On Fri, May 23, 2008 at 12:09 AM, David Francis [EMAIL PROTECTED] wrote:
 Greetings all

 I was looking at creating a little ZFS storage box at home using the 
 following SATA controllers (Adaptec Serial ATA II RAID 1420SA) on Opensolaris 
 X86 build

 Just wanted to know if anyone out there is using these and can vouch for 
 them. If not if there's something else you can recommend or suggest.

 Disk's would be 6*Seagate 500GB drives.

 Thanks

 David


 This message posted from opensolaris.org
 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Ian Collins
David Francis wrote:
 Greetings all

 I was looking at creating a little ZFS storage box at home using the 
 following SATA controllers (Adaptec Serial ATA II RAID 1420SA) on Opensolaris 
 X86 build

 Just wanted to know if anyone out there is using these and can vouch for 
 them. If not if there's something else you can recommend or suggest.

 Disk's would be 6*Seagate 500GB drives.

   
6 or more SATA slots are quite common on current motherboards, so if you
shop around, you may not need an add on card.

Ian

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Pascal Vandeputte
That 1420SA will not work, period. Type 1420sa solaris in Google and you'll 
find a thread about the problems I had with it.

I sold it and took the cheap route again with a Silicon Image 3124-based 
adapter and had more problems which now probably would be solved with the 
latest Solaris updates.

Anyway, I finally settled for a motherboard with an Intel ICH9-R and couldn't 
be happier (Intel DG33TL/DG33TLM, 6 SATA ports). No hassles and very speedy.

That Supermicro card someone else is recommending should also work without any 
issues, and it's really cheap for what you get (8 ports). Your maximum 
throuhput won't exceed 100MB/s though if you can't plug it in a PCI-X slot but 
resort to a regular PCI slot instead.

Greetings,

Pascal
 
 
This message posted from opensolaris.org
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Brian Hechinger
On Fri, May 23, 2008 at 12:47:18AM -0700, Pascal Vandeputte wrote:
 
 I sold it and took the cheap route again with a Silicon Image 3124-based 
 adapter and had more problems which now probably would be solved with the 
 latest Solaris updates.

I'm running a 3124 with snv81 and haven't had a single problem with it.
Whatever problems you ran into have likely been resolved.

Just my $0.02. ;)

-brian
-- 
Coding in C is like sending a 3 year old to do groceries. You gotta
tell them exactly what you want or you'll end up with a cupboard full of
pop tarts and pancake mix. -- IRC User (http://www.bash.org/?841435)
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Erik Trimble
Brian Hechinger wrote:
 On Fri, May 23, 2008 at 12:47:18AM -0700, Pascal Vandeputte wrote:
   
 I sold it and took the cheap route again with a Silicon Image 3124-based 
 adapter and had more problems which now probably would be solved with the 
 latest Solaris updates.
 

 I'm running a 3124 with snv81 and haven't had a single problem with it.
 Whatever problems you ran into have likely been resolved.

 Just my $0.02. ;)

 -brian
   
The Silicon Image 3114 also works like a champ, but it's SATA 1.0 only.  
It's dirt cheap (under $25), and you will probably need to re-flash the 
BIOS with one from Silicon Image's web site to remove the RAID software 
(Solaris doesn't understand it), but I've had nothing but success with 
this card (the re-flash is simple).


On a related note - does anyone know of a good Solaris-supported 4+ port 
SATA card for PCI-Express?  Preferably 1x or 4x slots...




-- 
Erik Trimble
Java System Support
Mailstop:  usca22-123
Phone:  x17195
Santa Clara, CA
Timezone: US/Pacific (GMT-0800)

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Brian Hechinger
On Fri, May 23, 2008 at 12:25:34PM -0700, Erik Trimble wrote:
 
 I'm running a 3124 with snv81 and haven't had a single problem with it.
 Whatever problems you ran into have likely been resolved.
 
 The Silicon Image 3114 also works like a champ, but it's SATA 1.0 only.  
 It's dirt cheap (under $25), and you will probably need to re-flash the 
 BIOS with one from Silicon Image's web site to remove the RAID software 
 (Solaris doesn't understand it), but I've had nothing but success with 
 this card (the re-flash is simple).

With the 3124 you don't even need to do the flash game, the 3124 is comletely
supported.

 On a related note - does anyone know of a good Solaris-supported 4+ port 
 SATA card for PCI-Express?  Preferably 1x or 4x slots...

The Silicon Image 3134 is supported by Solaris.

-brian
-- 
Coding in C is like sending a 3 year old to do groceries. You gotta
tell them exactly what you want or you'll end up with a cupboard full of
pop tarts and pancake mix. -- IRC User (http://www.bash.org/?841435)
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Tim
On Fri, May 23, 2008 at 2:36 PM, Brian Hechinger [EMAIL PROTECTED] wrote:

 On Fri, May 23, 2008 at 12:25:34PM -0700, Erik Trimble wrote:
  
  I'm running a 3124 with snv81 and haven't had a single problem with it.
  Whatever problems you ran into have likely been resolved.
  
  The Silicon Image 3114 also works like a champ, but it's SATA 1.0 only.
  It's dirt cheap (under $25), and you will probably need to re-flash the
  BIOS with one from Silicon Image's web site to remove the RAID software
  (Solaris doesn't understand it), but I've had nothing but success with
  this card (the re-flash is simple).

 With the 3124 you don't even need to do the flash game, the 3124 is
 comletely
 supported.

  On a related note - does anyone know of a good Solaris-supported 4+ port
  SATA card for PCI-Express?  Preferably 1x or 4x slots...

 The Silicon Image 3134 is supported by Solaris.



I'm looking on their site and don't even see any data on the 3134... this
*something new* that hasn't been released or?  The only thing I see is 3132.





 -brian
 --
 Coding in C is like sending a 3 year old to do groceries. You gotta
 tell them exactly what you want or you'll end up with a cupboard full of
 pop tarts and pancake mix. -- IRC User (http://www.bash.org/?841435)
 ___
 zfs-discuss mailing list
 zfs-discuss@opensolaris.org
 http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Brandon High
On Fri, May 23, 2008 at 12:43 PM, Tim [EMAIL PROTECTED] wrote:
 I'm looking on their site and don't even see any data on the 3134... this
 *something new* that hasn't been released or?  The only thing I see is 3132.

There isn't a 3134, but there is a 3124, which is a PCI-X based 4-port.

-B

-- 
Brandon High[EMAIL PROTECTED]
The good is the enemy of the best. - Nietzsche
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss


Re: [zfs-discuss] ZFS Project Hardware

2008-05-23 Thread Tim
On Fri, May 23, 2008 at 3:15 PM, Brandon High [EMAIL PROTECTED] wrote:

 On Fri, May 23, 2008 at 12:43 PM, Tim [EMAIL PROTECTED] wrote:
  I'm looking on their site and don't even see any data on the 3134... this
  *something new* that hasn't been released or?  The only thing I see is
 3132.

 There isn't a 3134, but there is a 3124, which is a PCI-X based 4-port.

 -B

 --
 Brandon High[EMAIL PROTECTED]
 The good is the enemy of the best. - Nietzsche


So we're still stuck the same place we were a year ago.  No high port count
pci-E compatible non-raid sata cards.  You'd think with all the demand
SOMEONE would've stepped up to the plate by now.  Marvell, cmon ;)

--Tim
___
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss