Matthew Crocker | 1 Jan 2008 17:43

PE1955 + RHEL5 best practices/documentation


Hello,

  I just racked my shiny new Dell Blade chassis with 10 PE1955s (2 x  
quad-core Xeon, 8 Gb RAM, 4 1Gb Ethernet).  The chassis has 4 Dell  
PowerConnect  switches.  I also have a 4.5 TB Dell MD3000i SAN array.   
All of the blades will run RHEL5 with Xen.
Gold 4 hour support on the whole deal :)

So, basically I have 10 systems, each with 4 GigE connections, 4 16  
port GigE switches, 1 GigE per server per switch.  I'd like to make  
this setup bullet proof,  I'm planning on 2 GigE switches for iSCSI   
to the MD3000i and 2 for data.    The Data switch will need to support  
about 5 802.1q VLANs (public, private, NFS, management, ...)

With 2 switches, 2 IP subnets connecting each blade to the M3000i can  
I load balance the traffic and get 2 GigE between the blade and the  
MD3000i?  4 GigE between the MD3000i and the chassis combined?

Is there a way I can bond 2 GigEs into a 2 gigabit EtherChannel across  
two PowerConnect switches?

Is there a way I can link the switches together inside the chassis or  
do I need to run an external cross over cable between them?

The MD3000i has a limit of 16 hosts.  Does that count as real physical  
servers or virtual servers?  i.e., can I run 4 Xen instances per blade  
and have them all run an iSCSI initiator?  Would that be 10 hosts or  
40 from the MD3000i point of view?

(Continue reading)

Brian A. Seklecki | 1 Jan 2008 19:51

RE: 10 gigabit ethernet - Add on question

On Mon, 31 Dec 2007, Robert D. Holtz - Lists wrote:

> I would also be curious which OS flavor and hardware combination you would
> need to actually put a sustained 10gbps onto the wire.

Meh.  NetBSD/amd64 runs fine on the 2950/2850.  We set the record using 
Intel 10gb back in the day (2004):

   http://bsd.slashdot.org/article.pl?sid=04/05/03/2235255&tid=95

TCP/IP had its development origins in *BSD and in 2008 we still have the 
best stack out there.

~BAS

(Run for the hills.)

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

Edgar Matzinger | 2 Jan 2008 01:21
Picon

Possible to use a Pinnacle PCTV Hybrid Pro PCI in a PE4600?

Hello list,

  I recently got me a PE4600 with the intention to run it as a MythTV
(backend) server. Now I know the PE4600 has a 32bit/33MHz PCI slot,
which I can use to put in a Hauppauge PVR 250... But the beast has 7
slots more. These are 64bit/66MHz PCI-X slots  <at>  3.3v. The question is
can I use a PCI rev 2.3 card (like the Pinnacle PCTV Hybrid Pro PCI)
in such a slot?

Many thanks, cu l8r, Edgar.
--

-- 
       \|||/
       (o o)                                             Just curious...
----ooO-(_)-Ooo---------------------------------------------------------

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

Jimmy Christensen | 2 Jan 2008 09:40
Picon
Favicon

Re: Redundant NFS storage setup (part 3) : The disappointing PERC5/E

Hi,

I would suggest testing with a stripe size of 64kb instead of 128kb. We 
are currently testing our HP servers with a P800 with similar 
dissapointment, but found out that anything above a stripe size of 64kb 
absolutely kills the performance (ofcourse no one would tell us that). I 
believe it's based on a similar chip that the PERC5/E has.

Jimmy

Matthias Saou wrote:
> Hi,
>
> I'm in the final stages of the file server setup I've already
> mentioned twice on the list.
>
> Right now, I'm still evaluating the production hardware :
> - PowerEdge 1950 III with 16GB RAM and dual Quad Core E5410
> - PowerVault MD1000 with 15 1TB SATAII drives
>
> I initially configured the PERC5/E with a single RAID5 array of 14
> disks and 1 global hot spare. I was VERY disappointed with the
> overall I/O performance of the volume. So I tried with RAID10. Again,
> performance was mediocre at best, including simple reads.
>
> A few numbers. I use RHEL5 x86_64 and XFS for the filesystem :
>
> Non cached 10GB file read :
>  * RAID5 : 255MB/s
>  * RAID10 : 224MB/s (???)
(Continue reading)

Harald_Jensas | 2 Jan 2008 11:05
Picon
Favicon

RE: PE1955 + RHEL5 best practices/documentation

> -----Original Message-----
> From: linux-poweredge-bounces <at> dell.com [mailto:linux-poweredge-
> bounces <at> dell.com] On Behalf Of Matthew Crocker
> Sent: 01 January 2008 17:44
> To: linux-poweredge-Lists
> Subject: PE1955 + RHEL5 best practices/documentation
> 
> 
> Hello,
> 
>   I just racked my shiny new Dell Blade chassis with 10 PE1955s (2 x
> quad-core Xeon, 8 Gb RAM, 4 1Gb Ethernet).  The chassis has 4 Dell
> PowerConnect  switches.  I also have a 4.5 TB Dell MD3000i SAN array.
> All of the blades will run RHEL5 with Xen.
> Gold 4 hour support on the whole deal :)
> 
> So, basically I have 10 systems, each with 4 GigE connections, 4 16
> port GigE switches, 1 GigE per server per switch.  I'd like to make
> this setup bullet proof,  I'm planning on 2 GigE switches for iSCSI
> to the MD3000i and 2 for data.    The Data switch will need to support
> about 5 802.1q VLANs (public, private, NFS, management, ...)
> 
> With 2 switches, 2 IP subnets connecting each blade to the M3000i can
> I load balance the traffic and get 2 GigE between the blade and the
> MD3000i?  4 GigE between the MD3000i and the chassis combined?

Yes. There is MultiPath I/O Software included on the MD3000i Resource CD.

Documentation is available here:
http://support.euro.dell.com/support/edocs/systems/md3000i/en/IG/PDF/IGbk0HR.pdf
(Continue reading)

Mike Hanby | 2 Jan 2008 16:20
Picon
Favicon

RE: Install OpenManage 5.3.0 Managed Node fails with no errors

Howdy Patrick, thanks for the reply. Another user pointed this out off list as well, it completely escaped me that CentOS wasn't a supported platform.

 

Thanks for the suggestion, I actually modified the script, although temporarily replacing /etc/redhat-release would have been easier in retrospect :-)

 

From: Patrick_Boyd <at> Dell.com [mailto:Patrick_Boyd <at> Dell.com]
Sent: Wednesday, January 02, 2008 09:16
To: Mike Hanby; linux-poweredge <at> lists.us.dell.com
Subject: RE: Install OpenManage 5.3.0 Managed Node fails with no errors

 

Have you created a /etc/redhat-release file containing “Red Hat Enterprise Linux ES release 4 (Nahant)”?

 

From: linux-poweredge-bounces <at> dell.com [mailto:linux-poweredge-bounces <at> dell.com] On Behalf Of Mike Hanby
Sent: Monday, December 31, 2007 9:43 AM
To: linux-poweredge-Lists
Subject: Install OpenManage 5.3.0 Managed Node fails with no errors

 

Howdy,

 

I've installed OpenManage 5.3.0 (Managed Node) on several PE 2850 and 2950's using the installation script in the file: OM_5.3.0_ManNode_A00.tar.gz

 

I'm unable to install it one server, however.

 

I run setup.sh, the script runs for 15 seconds or so and then returns to the command prompt.

 

This system is running CentOS 4.4 x86_64 with kernel 2.6.9-42.0.2.ELsmp

 

 

I enabled logging (export OMIDEBUG=1) and all I see is the following:

 

# cat /var/log/srvadmin.log

 

------------------------------------------------------------------------

Server Administrator Custom Install Start

Mon Dec 31 09:41:28 CST 2007

------------------------------------------------------------------------

 

------------------------------------------------------------------------

Server Administrator Custom Install Stop

Mon Dec 31 09:41:29 CST 2007

------------------------------------------------------------------------

 

------------------------------------------------------------------------

Server Administrator Custom Install Stop

Mon Dec 31 09:41:34 CST 2007

------------------------------------------------------------------------

 

Any suggestions?

 

Thanks, Mike

 

-----

Mike Hanby

Information Systems Specialist II

School of Engineering Dean's Office

University of Alabama at Birmingham

 

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq
Patrick_Boyd | 2 Jan 2008 16:10
Picon
Favicon

RE: Dell NX1950 NAS Question

I worked on this project, and have one of these boxes serving production
duty as a file server. So far I've been fairly impressed with it. I have
all the shares setup as both NFS and SMB so I can use native protocols
on Windows and Linux. So far the NFS seems to be much faster and
reliable than I have found CIFS to be.

There is some setup overhead to deal with, which is mapping the
usernames/groups or setting up a mapping server, but once that is done I
think you'll be quiet happy with the NX1950 as a file server.

-----Original Message-----
From: linux-poweredge-bounces <at> dell.com
[mailto:linux-poweredge-bounces <at> dell.com] On Behalf Of Rogers, Jamie
Sent: Friday, December 21, 2007 1:01 PM
To: linux-poweredge-Lists
Subject: Dell NX1950 NAS Question

Does anyone have any experience with Dells NX1950 NAS system?  I'm
looking for another storage solution and I'm looking for some anecdotal
reports.  I have a mixed Solaris, Linux RHEL, AIX, HP-UX, and CIFS,
environment that serves a number of HPC for simulations.  I currently
use Netapp's to serve the data but I'm getting maxed out there.  I
really don't like the thought of using Windows Unified Data Storage
Server to serve up files to Unix through NFS but I'm willing to listen
to what others have to say about their experiences.  I know that Windows
2003R2 is not certified for a number of OS's that I support, but I'm not
sure about Windows UDSS.  I'll spend some time reading on Mirosoft's
site but I'm still interested in what other users have to say.  

Thanks in advance,
Jamie 

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

Sam Flory | 2 Jan 2008 18:38
Favicon

Re: 10 gigabit ethernet

Hall, James F CTR USAF ASC/ASC/HP wrote:
> Which card(s) are best supported by Dell and Dell's Linux team for 10Gb
> Ethernet support?
>
> I am also looking for possible implementation of FCoE, Virtualization
> support, and PCI-E, PCI-X, and low profile solutions.
>   

  The 2950-III, and 1950-III both have Intel 10GbE controller options. 
Last I checked Dell supports the latest RHEL 5 release on it.  Also RHEL
4.(what is the latest) should work.  As should SLES, and any up to date
version of linux.  Of course the 10G card will x2-3 the price of the
system.  The intel 10G is in the mainline kernel.

See tech specs:
http://www.dell.com/content/products/productdetails.aspx/pedge_2950_3

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq

Patrick_Boyd | 2 Jan 2008 16:15
Picon
Favicon

RE: Install OpenManage 5.3.0 Managed Node fails with no errors

Have you created a /etc/redhat-release file containing “Red Hat Enterprise Linux ES release 4 (Nahant)”?

 

From: linux-poweredge-bounces <at> dell.com [mailto:linux-poweredge-bounces <at> dell.com] On Behalf Of Mike Hanby
Sent: Monday, December 31, 2007 9:43 AM
To: linux-poweredge-Lists
Subject: Install OpenManage 5.3.0 Managed Node fails with no errors

 

Howdy,

 

I've installed OpenManage 5.3.0 (Managed Node) on several PE 2850 and 2950's using the installation script in the file: OM_5.3.0_ManNode_A00.tar.gz

 

I'm unable to install it one server, however.

 

I run setup.sh, the script runs for 15 seconds or so and then returns to the command prompt.

 

This system is running CentOS 4.4 x86_64 with kernel 2.6.9-42.0.2.ELsmp

 

 

I enabled logging (export OMIDEBUG=1) and all I see is the following:

 

# cat /var/log/srvadmin.log

 

------------------------------------------------------------------------

Server Administrator Custom Install Start

Mon Dec 31 09:41:28 CST 2007

------------------------------------------------------------------------

 

------------------------------------------------------------------------

Server Administrator Custom Install Stop

Mon Dec 31 09:41:29 CST 2007

------------------------------------------------------------------------

 

------------------------------------------------------------------------

Server Administrator Custom Install Stop

Mon Dec 31 09:41:34 CST 2007

------------------------------------------------------------------------

 

Any suggestions?

 

Thanks, Mike

 

-----

Mike Hanby

Information Systems Specialist II

School of Engineering Dean's Office

University of Alabama at Birmingham

 

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq
Karl Zander | 2 Jan 2008 20:13
Favicon

Troubleshooint PowerVault 100T Travan 40 tape drive

We have a PE 1420 with a PowerVault 100T Travan 40 tape drive 
installed.  dmesg reports it as

hdd: Seagate STT3401A, ATAPI TAPE drive
ide-tape: hdd <-> ht0: Seagate STT3401A rev 310A

The documentation for the drive says there is should be a Tape 
Diagnostic Utility available. 
 http://support.dell.com/support/edocs/stor%2Dsys/travan40/en/trouble.htm

I have looked at the downloads at http://support.dell.com and can't 
find it.  Anyone know where we can get it?

--Karl

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge <at> dell.com
http://lists.us.dell.com/mailman/listinfo/linux-poweredge
Please read the FAQ at http://lists.us.dell.com/faq


Gmane