Re: HP ProLiant DL320 v. Sun Fire V125

Previous thread: Re: Network Time Synchronization using timed or ntpd or a Combination? by Christian Weisgerber on Tuesday, October 23, 2007 - 2:28 pm. (1 message)

Next thread: Re: About Xen: maybe a reiterative question but .. by Henning Brauer on Wednesday, October 24, 2007 - 1:18 am. (84 messages)
From: evo
Date: Tuesday, October 23, 2007 - 10:51 pm

Hello, misc.

I'm choosing firewall/proxy/mail-gateway hardware running (of course) 
OpenBSD for medium office and my shortlist is:
(a) HP ProLiant DL320 and (b) Sun Fire V125

Price for both servers is more or less close.
Personally I'd prefer to buy Sun, just for opportunity to play with 
non-i386 arch :) but according to http://www.openbsd.org/sparc64.html:
"Sun refused access to the neccessary documentation for the (very 
bizzare) host bridge in the UltraSPARC III machines, so a few years were 
lost before some reverse engineering figured out the changes in these 
machines. OpenBSD 4.0 has thus been the first release to ship with 
support for the UltraSPARC III based machines.
And since there are always little bits missing, work is continuing..."

This confusing me a bit. Does OpenBSD/sparc64 performs well? I need to 
run pf, squid, postfix, amavisd-new, spamassassin, clamav - will this 
work on Sun as good as on my current intel-based-office-pc-router?

Which one's hardware better supported? Internal NIC's, SCSI 
(Sun)/SATA(HP) controllers?

I'd like to hear from people who use these machines in production - 
what's good, what's bad, what's evil.

And if you had to choose - what would be your choice today?

Thanks for attention.

--
Yuri A. Spirin
Volgograd Mechanical Engineering Company "VgTZ"
Systems Administrator

From: Boris Goldberg
Date: Wednesday, October 24, 2007 - 11:41 am

Hello evo,

Wednesday, October 24, 2007, 12:51:13 AM, you wrote:

e> I'm choosing firewall/proxy/mail-gateway hardware running (of course)
e> OpenBSD for medium office and my shortlist is:
e> (a) HP ProLiant DL320 and (b) Sun Fire V125

  I'm   upgrading   my  servers/firewalls  to HP ProLiant DL320 G5, and the
experience...  isn't  easy.  First  of  all you need to allow acpi in an MP
kernel,  otherwise it's slow and unstable (it's disabled by default and not
really documented).

  Then you have couple more issues I couldn't resolve yet:

  Fists - uhci (uhci4 in my case) giving an error during boot and shutdown:

OpenBSD 4.2-stable (GENERIC) #1: Thu Oct 18 12:35:10 CDT 2007
    root@build.twopoint.com:/usr/src/sys/arch/i386/compile/GENERIC
cpu0: Intel(R) Pentium(R) D CPU 3.00GHz ("GenuineIntel" 686-class) 3.01 GHz
cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,EST,CNXT-ID,CX16,xTPR
real mem  = 1071640576 (1021MB)
avail mem = 1028595712 (980MB)
mainbus0 at root
bios0 at mainbus0: AT/286+ BIOS, date 12/31/99, BIOS32 rev. 0 @ 0xf0000, SMBIOS rev. 2.3 @ 0xee000 (47 entries)
bios0: vendor HP version "W04" date 04/06/2007
bios0: HP ProLiant DL320 G5
pcibios0 at bios0: rev 3.0 @ 0xf0000/0x2000
pcibios0: PCI BIOS has 7 Interrupt Routing table entries
pcibios0: PCI Interrupt Router at 000:31:0 ("Intel 82801GB LPC" rev 0x00)
pcibios0: PCI bus #7 is the last bus
bios0: ROM list: 0xc0000/0xb000 0xcc400/0x1000 0xcd400/0x1000 0xce400/0x3400! 0xe6000/0x2000!
acpi at mainbus0 not configured
ipmi0 at mainbus0: version 2.0 interface KCS iobase 0xca2/2 spacing 1
cpu0 at mainbus0
pci0 at mainbus0 bus 0: configuration mode 1 (no bios)
pchb0 at pci0 dev 0 function 0 "Intel E7230 MCH" rev 0xc0
ppb0 at pci0 dev 1 function 0 "Intel E7230 PCIE" rev 0xc0
pci1 at ppb0 bus 1
ppb1 at pci0 dev 28 function 0 "Intel 82801GB PCIE" rev 0x01
pci2 at ppb1 bus 2
ppb2 at pci2 dev 0 function 0 "ServerWorks ...
From: Kai Mosebach
Date: Wednesday, October 31, 2007 - 6:57 am

Hi,

We run quite fine here with 4.2-current from today on a DL320G5, after:

enabling write cache in the HP Bios !
enabling amd64 bsd.mp
enabling acpi
enabling write cache for wd0 in the system with:

# atactl wd0 writecacheenable

Before we had horrible 2MByte write speed, now we have 67MByte.

The bge interfaces also seem to run fine.

cheers Kai

manam:~ # dmesg
OpenBSD 4.2-current (GENERIC.MP) #1436: Mon Oct 29 23:01:38 MDT 2007
     deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP
real mem = 2145415168 (2046MB)
avail mem = 2071965696 (1975MB)
mainbus0 at root
bios0 at mainbus0: SMBIOS rev. 2.3 @ 0xee000 (47 entries)
bios0: vendor HP version "W04" date 03/10/2007
bios0: HP ProLiant DL320 G5
acpi0 at mainbus0: rev 2
acpi0: tables DSDT FACP SPCR MCFG HPET SPMI APIC
acpitimer at acpi0 not configured
acpimadt0 at acpi0 addr 0xfee00000: PC-AT compat
cpu0 at mainbus0: apid 0 (boot processor)
cpu0: Intel(R) Xeon(R) CPU 3040 @ 1.86GHz, 1866.98 MHz
cpu0: 
FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,EST,TM2,CX16,xTPR,LONG
cpu0: 2MB 64b/line 8-way L2 cache
cpu0: apic clock running at 266MHz
cpu1 at mainbus0: apid 1 (application processor)
cpu1: Intel(R) Xeon(R) CPU 3040 @ 1.86GHz, 1866.74 MHz
cpu1: 
FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,EST,TM2,CX16,xTPR,LONG
cpu1: 2MB 64b/line 8-way L2 cache
ioapic0 at mainbus0 apid 8 pa 0xfec00000, version 20, 24 pins
acpiprt0 at acpi0: bus 1 (IP2P)
acpiprt1 at acpi0: bus 4 (PCXS)
acpiprt2 at acpi0: bus 3 (PCXA)
acpiprt3 at acpi0: bus 2 (ICHE)
acpiprt3: no apic found for irq 47
acpiprt4 at acpi0: bus 7 (IPE4)
acpiprt5 at acpi0: bus 10 (PTA0)
acpiprt6 at acpi0: bus 0 (PCI0)
acpicpu at acpi0 not configured
acpicpu at acpi0 not configured
acpicpu at acpi0 not configured
acpicpu at acpi0 not configured
acpicpu at ...
From: Boris Goldberg
Date: Sunday, November 4, 2007 - 10:21 am

Hello Kai,

  Thank you very much for the reply. It's helpful.

Wednesday, October 31, 2007, 8:57:53 AM, you wrote:

KM> We run quite fine here with 4.2-current from today on a DL320G5, after:

KM> enabling write cache in the HP Bios !

  It  looks  like  that  BIOS  "write cache" settings don't change anything
(atactl does it).

KM> enabling amd64 bsd.mp

  Your CPU is Xeon, mine is Pentium D. Don't think amd64 will work for me.

KM> enabling acpi

  How exactly do you do it?
  Mine acpi-related lines are

#option         ACPIVERBOSE
#option         ACPI_ENABLE

acpi0           at mainbus?
acpitimer*      at acpi?
#acpihpet*      at acpi?
#acpiac*        at acpi?
#acpibat*       at acpi?
#acpibtn*       at acpi?
acpicpu*        at acpi?
#acpidock*      at acpi?
acpiec*         at acpi?
acpiprt*        at acpi?
acpitz*         at acpi?

  Do I need to uncomment "options" or they are active by default anyway? Is
there any documents about it?

KM> enabling write cache for wd0 in the system with:
KM> # atactl wd0 writecacheenable

  Where  do  you  put  these  command?  For now I just ran it manually (and
tested  the  result).  I  think  it makes sense to activate the write cache
before  checking (and possibly recovering) RAIDframe devices, but rc.secure
and  rc.local  are  being called after that. Is it a good idea to put these
atactl commands to /etc/rc right before "#Configure ccd devices" line?

KM> Before we had horrible 2MByte write speed, now we have 67MByte.

  I'm  getting  an about 16 times speed increase on copying a 1.2 gig file.
Is there any performance tests for the OpenBSD, BTW?

KM> The bge interfaces also seem to run fine.

  Have  you  tried to boot with a network cable unplugged and than plug it?
My  bge*  (on  two  computers so far) detects a media of 10 megabit in that
case  (ifconfig  down/up  makes  it to detect the right media - 100 or 1000
megabit). em* devices don't have that (minor?) bug.

KM> "Compaq iLO" rev 0x03 at pci6 ...
From: Kai Mosebach
Date: Thursday, December 6, 2007 - 8:29 am

Hi,


its already in the default kernel, not sure if its enabled by default.

# config -ef /bsd.mp
...
ukc> enable acpi

i ran it manually as well, and i guess you're right, even though i would 
not touch /etc/rc, since its part of the core system and i guess its not 
supposed to be changed by the "user"
dunno if there is a sysctl and or other way to enable it at boot time, 

maybe try to disable it in the kernel config

# config -ef /bsd.mp
...
ukc> disable uhci

we dont use it. sorry.

Best Kai

Previous thread: Re: Network Time Synchronization using timed or ntpd or a Combination? by Christian Weisgerber on Tuesday, October 23, 2007 - 2:28 pm. (1 message)

Next thread: Re: About Xen: maybe a reiterative question but .. by Henning Brauer on Wednesday, October 24, 2007 - 1:18 am. (84 messages)