September 2009 Archives

Part 1 of my Solaris Fiberchannel notes

| | Comments (1) | TrackBacks (0)
wwnnwwpnhost.pngWhat's a WWPN? A WWNN? What's an F_port? How my disks appear in Solaris?

As many readers asked me to give an introduction to fiberchannel usage in Solaris I started a little course for you.

Lesson 1 are fundamentals of the fiberchannel protocol - not really much, but only  the really needed addressing and network stuff.

It begins here:

I love to have feedback from you, and please ask if you don't understand something or when my language is too bad or too simple.

One strong remark: I am a german citizen writing in english to have a greater audience. Please don't hesitate to correct my faults in english. I'll be happy about that!
A step-by-step instruction manual how to install RHEL 5 using a paravirtual SCSI boot device in VMware is described here:

Comments are welcome!

Sun MPxIO Storage / scsi_vhci: Solaris 10 vs OpenSolaris

| | TrackBacks (0)
In Solaris 11 (SXCE, OpenSolaris) the format of the configuration file "scsi_vhci.conf" has changed. As a starter let's remember how MPxIO for storage works (here Fiberchannel is used as an example, the same applies to multiple iSCSI links, just replace the fc driver layer with ethernet/ip/tcp/iscsi):

This is the example with two distinct fiberchannel ports present (multiport FC adapter or distinct FC adapters).

rtc! Sun xVM: Wrong timezone offset in PV DomU?

| | TrackBacks (0)
Running paravirtualized Linux Kernels as DomU on Sun xVM/Xen 3.4.2 I had the "little" problem that timezone offsets are not correct. My machines are located in Germany so the Timezone in Linux is "Europe/Berlin". The result:

OpenSolaris Dom0: Tuesday, September 22, 2009  7:06:38 AM UTC
Linux SLES 11 DomU: Tue Sep 22 11:06:38 CEST 2009

"CEST" means Central European Summer Time, and the offset should be two hours, not four!

The solution:
Dom0's real time (hardware) clock has to be set to UTC!
Using Linux as Dom0 this would be a simple  "hwclock --utc" call. On OpenSolaris the command is:

rtc -z UTC

The result will be written permanently in /etc/rtc_config.
Don't forget to do a
svcadm disable ntp; ntpdate ...yourtimeserver...; svc enable ntp

DomU Linux machines are immediately changing to the right time (9:06 CEST in the example above).

How storage works in VMWare ESX?

| | TrackBacks (0)
As opposed to the Xen/Sun xVM storage virtualization is done in the ESX Kernel (the simple case where no dedicated hardware storage is reserved and defined as passthrough for the virtual machine):


How storage works in Sun xVM/Xen?

| | TrackBacks (0)
In the HVM case it is a little bit like this (simplified):

Redhat Enterprise 5 Server in Xen!

| | TrackBacks (0)
Yes we know, Redhat Inc. has its own virtualization technique named KVM but nevertheless it installs just painless in a Xen environment.

I spent 15 minutes with an interactive install in a paravirtualized Sun xVM/Xen domain and this is very fast.

I described the steps in my little article: Installing Redhat Enterprise Linux 5 on Sun xVM

It appears in the Redhat network as "Para-virtualized" which is 100% correct:


Virtualization techniques and approaches

| | TrackBacks (0)
virtualization-hypervisor-example-60.pngI began to compare some different approaches to virtualization available and to visualize them with graphical pictograms.

I am appreciating your feedback whether I achieved that goal or not.

The following articles are available so far:

  1. Why virtualization?
  2. Userspace-based virtualization (the easy way)
  3. Xen: Hypervisor-based virtualization
  4. ESX: Hypervisor-based virtualization
  5. Solaris Zones: A sharing approach
  6. Hybrid methods: KVM

ESX 4 (VMWare): SUSE Linux with paravirtual SCSI controller

| | TrackBacks (0)
[Remark: With Redhat Enterprise Server 5 it is a little bit more difficult, creation of the initial ramdisk image is not so straightforward: Click here for the redhat howto]

works out of the box:

SCSI subsystem initialized
VMware PVSCSI driver - version
pvscsi 0000:03:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
pvscsi: found VMware PVSCSI rev 1 on bus 3:slot 0:func 0
pvscsi 0000:03:00.0: setting latency timer to 64
pvscsi: enabled MSI-X
scsi0 : VMware PVSCSI storage adapter rev A, 256 reqs (8 pages), 1024 cmps (8 pages), cmd_per_lun=64
scsi 0:0:0:0: Direct-Access     VMware   Virtual disk     1.0  PQ: 0 ANSI: 2
 sda: sda1 sda2 sda3
if you do the following steps:

vSphere 4 ESX vs Sun xVM Xen 3.4.2: Windows 2003 SP2

| | Comments (2) | TrackBacks (0)
I am not a Windows guy, I admit, but in our data center there are many Windows 2003 Server installations (32bit) and we are going to virtualize them.

Management's choice is VMWare vSphere 4 (ESX 4) but nevertheless I wanted to know how Sun xVM/Xen behaves compared to ESX. Forget all that nice Management tools offered by vmware, the only thing I was interested in is the performance feeling when actually using them.

The picture at the right (click on it to enlarge) shows the hardware list of the two "brothers": To the left you'll see the xVM DomU with paravirtualized disk and network drivers, to the right there's the vSphere ESX virtual machine with a paravirtualized SCSI disk device (and an IDE boot disk, you cannot boot in ESX from a paravirtualized disk).


| | Comments (4) | TrackBacks (0)
How old were you 1987?
10srmodii.JPGI was 16 years old and I had my Commodore 64 with a 1541 floppy disk and 170 kilobytes was a big storage for me, a guy used to cassette tapes to store computer data.
In real business, hard disks already carried hundreds of megabytes and were big - yes big. One of these "disks" was brought in by my esteemed collegue who got it from a bank some years ago where these platters ran for 12 years without any problems.

I took some pictures of this machine and I am pleased to show them to you.

Manufacturing date is 1987, the "mod I" had 200 Megabytes, but this is a "mod II", I think it had 300 Megabytes.

This hard disk can now be visited at the computing center of the University of Konstanz, central hall in building V, 5th floor.

Assigning CPUs works even in Domain-0!

| | TrackBacks (0)
Just a small note:

It was just a test, but it worked!

You can control the number of CPU cores Dom0 uses (named "Domain-0" in the Sun xVM environment).

This command:

# virsh setvcpus Domain-0 2

results in that:

Sep  2 18:35:16 pgt01 unix: cpu0: externally initiated on-line
Sep  2 18:35:17 pgt01 unix: cpu1: externally initiated on-line
Sep  2 18:35:17 pgt01 unix: cpu2: externally initiated powered-off
Sep  2 18:35:18 pgt01 unix: cpu3: externally initiated powered-off

OS is OpenSolaris snv_121, which was bfu'd from snv_118.

(Setting the memory use and max memory use via virsh setmem and virsh setmaxmen do also work as expected, the system begins to swap out pages and/or flushes zfs cache data on disk).

No more Yahoo IM for old clients

| | TrackBacks (0)
Yahoo closed its messaging servers yesterday for older clients using older versions of the protocol.

Kopete of KDE 3.5 is one of these.

The trend to force users to use the proprietary bloated clients is getting stronger.

December 2015

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    


This blog is owned by:

Pascal Gienger
J├Ągerstrasse 77
8406 Winterthur

Google+: Profile
YouTube Channel: pascalgienger