7 dic. 2012

Memory and CPU hotplug on Ubuntu Guests and VMWare ESXi hosts (including vCloud)

Introduction

With some cloud projects in mind and some virtual hardware estimations I was curious to see if it's possible to hot plug virtual memory and CPU to Ubuntu guests. Here is what I discovered.

  • Infrastructure: vCloud Director on Acens.com (VMWare ESXi)
  • Guest: Ubuntu server 10.04 (kernel 2.6.32-45-server #100-Ubuntu SMP Wed Nov 14 11:02:27 UTC 2012 x86_64 GNU/Linux)
    • virtual HW version: 7
    • vCPU: 1
    • vMemory: 1 GB

Hot add virtual CPUs

Check current CPUs:

# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model  : 37
model name : Intel(R) Xeon(R) CPU           L5640  @ 2.27GHz
stepping : 1
cpu MHz  : 2266.747
cache size : 12288 KB
fpu  : yes
fpu_exception : yes
cpuid level : 11
wp  : yes
flags  : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc up arch_perfmon pebs bts rep_good xtopology tsc_reliable nonstop_tsc aperfmperf pni pclmulqdq ssse3 cx16 sse4_1 sse4_2 popcnt aes hypervisor lahf_lm arat
bogomips : 4533.49
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management:

Now, add the virtual CPUs directly from the vCloud control panel (from the hardware tab of the VM properties). Initially, these new CPUs are offline. To active each new CPU we have to:

echo 1 > /sys/devices/system/cpu/cpuX/online

where X is the number of the new CPU.

Check that new CPUs are online:

# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model  : 37
model name : Intel(R) Xeon(R) CPU           L5640  @ 2.27GHz
stepping : 1
cpu MHz  : 2266.747
cache size : 12288 KB
fpu  : yes
fpu_exception : yes
cpuid level : 11
wp  : yes
flags  : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc up arch_perfmon pebs bts rep_good xtopology tsc_reliable nonstop_tsc aperfmperf pni pclmulqdq ssse3 cx16 sse4_1 sse4_2 popcnt aes hypervisor lahf_lm arat
bogomips : 4533.49
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management:

processor : 1
vendor_id : GenuineIntel
cpu family : 6
model  : 37
model name : Intel(R) Xeon(R) CPU           L5640  @ 2.27GHz
stepping : 1
cpu MHz  : 2266.747
cache size : 12288 KB
fpu  : yes
fpu_exception : yes
cpuid level : 11
wp  : yes
flags  : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc up arch_perfmon pebs bts rep_good xtopology tsc_reliable nonstop_tsc aperfmperf pni pclmulqdq ssse3 cx16 sse4_1 sse4_2 popcnt aes hypervisor lahf_lm arat
bogomips : 4533.49
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management:

You can use the script provided in this article:

http://communities.vmware.com/docs/DOC-10493

This enables automatically all new vCPUs.

Hot add virtual memory

This operations is analogue to one described above, but the kernel module that detects the new memory is not automatically loaded (at least in my Ubuntu install), so we have to do it manually:
# modprobe acpi_memhotplug

Let's check current memory first:

# free -m
             total       used       free     shared    buffers     cached
Mem:           997        112        885          0         10         42
-/+ buffers/cache:         59        937
Swap:         2015          0       2015

Add the memory through vCloud (or VMWare vSphere Infrastructure Cliente - VCLI). Although the kernel will have detected the new memory, it's still not available. We have to enable it in the same way we did it before with virtual CPUs:

echo 1 > /sys/devices/system/memory/memoryX/online

where X is the number of the new memory block.

After the memory blocks have been enabled, we can check the available memory:

# free -m
             total       used       free     shared    buffers     cached
Mem:          1893        141       1752          0         10         42
-/+ buffers/cache:         88       1805
Swap:         2015          0       2015

As with CPU hot plug, you can use the script provided in this article:

http://communities.vmware.com/docs/DOC-10492

This enables automatically all new memory blocks.

Epilogue

It's easy to add new resources on the fly but it may not be that easy to remove them.

30 nov. 2012

Compile Nginx with custom OpenSSL in Ubuntu 10.04

The problem

We need Server Name Indication (SNI) for nginx, but the OpenSSL version included in Ubuntu 10.04 does not support it.

Edit: I just noticed that I was wrong. SNI  *IS* supported by the version of openssl provided with Ubuntu 10.04:  libssl0.9.8-7ubuntu8.13 

The solution

  • Download, build and install a recent version of OpenSSL
  • Compile nginx against this version of OpenSSL
  • Do not replace the system openssl
Here are the steps I took:

# install dependencies
sudo apt-get install build-essential libpcre3-dev libxml2-dev libxslt1-dev
cd
mkdir -p src; cd src
wget http://www.openssl.org/source/openssl-1.0.1c.tar.gz
tar xzf openssl*
cd openssl*
./config shared zlib-dynamic
make
# This installs everything in /usr/local/ssl
sudo make install
wget http://nginx.org/download/nginx-1.2.5.tar.gz
tar xzf nginx*
cd nginx*
# edit auto/lib/openssl/conf manually or use sed
sed -i -e 's|\.openssl/||' auto/lib/openssl/conf
./configure --with-openssl=/usr/local/ssl --with-http_ssl_module
make
# test if SNI is displayed
./objs/nginx -V
# if everithing is ok, install
sudo make install

References




10 oct. 2012

Restoring iPhone under Ubuntu 12.04 / VirtualBox / WinXP / iTunes

Today, I bricked my iPhone and here goes my experience.

My plataform: Ubuntu 12.04/amd64, Oracle VirtualBox 4.2.0 / WinXP SP 2, iTunes 10.x

Some block posts reported permission problems of USB device files and adviced starting VirtualBox as root. And so I did. But the restoration / update process stuck.

Problem 1: VirtualBox does not recognize when a new USB device is plugged in. This may be due to starting VirtualBox as root, but I'm not sure, so stop or hibernate guest OS, do /etc/init.d/vboxdrv restart, and run guest again.

Problem 2: VirtualBox stops anywhere in the update process. There seems to be a problem with constant USB device switching (iPhone recovery mode), etc. It was just a matter of disconnecting the USB device with help of VirtualBox USB menu. It gets automatically reconnected and the update process continues.


I'll see if I find a cleaner way to do this.