As part of my nmon for Linux open source project, I often get asked questions and I have selected one here that comes up often. The nmon user asks is nmon only reporting 48 CPUs when he knows the machine has 24 CPU and they are 6 core each? The first answer was - Why the heck are you using a 4 years out of date nmon for Linux version?? But once we get passed that shocker, we look in to the details nmon has collected.   From the nmon data in the BBB Section, we have the /proc/cpuinfo file.  It is massive and has large amounts... [More]

Over the Xmas break I have been busy improving nmon and this is a summary. Below is a summary with pictures: 1 New Performance Statistics GPU First for the OpenPOWER people using the Nvidia GPUs within the S822LC (Firestone) machines we have the Graphical Processor Units and a previous AIXpert Blog covered the details but below is a little reminder: These are the stats on-screen GPU MHz CPU and Memory Utilisation Temperature in C Electrical power in Watts This is also saved to the nmon output file.   MHz Below we now noticed the CPU... [More]

The High Performance guys at IBM thought it would be good to add Nvidia K80 GPUs stats to my nmon for Linux open source tool and who am I to complain about that?  My new best friend :-) Manish Modani, gave me remote access to an internal machine in Austin, Texas from London over the IBM network and once the IBM firewall allowed me on the Austin lab network  I was logged in.   Then the fun started as I know nothing about these GPU devices.  First we came across a NVidia command that outputs a lot of stats on... [More]

I am upgrading my VIOS to 2.2.4.10 which includes SSP5 Tiers functions and LU grow functions and lots of other improvements To keep track of the nodes that still need to upgraded you can use the VIOS cluster -status -verbose command but the details fly up the screen and ate not visible without scrolling back and are gone and I really don't what to see all those massive hexadecimal internal Id's etc.   So here is a quick one line script "ncluster" /usr/ios/cli/ioscli cluster -status -verbose -fmt : -field... [More]

In the OpenPOWER IBM S812LC - First Look we had a look at the basic machine but in this blog we now have 12 disks to add and a RAID SATA adapter to drive them. I have made an update with new information at the end of this Blog in red. Here is a repeat of the first picture but highlighting the details for supporting the twelve 3.5 inch Front Disk Bays: Note: The three blue SATA cables are provided with the basic system but not connected to anything, of course, until now.   Below the delivery unpacked... [More]

Tags:  power8 redhat s812lc ubuntu tyan powerkvm openpower

If you just want the S812LC pictures, you should scroll down a bit ...     For reasons I do not understand there are very few picture of this new machine from IBM.  So I want to fix that in this blog entry and as a picture says 1000 words this will hopefully explain a lot about the machine - the equivalent of 37,000 words!   First, thing you need to know is that it is not in the regular IBM Power Systems series of machines. It is designed and developed by OpenPOWER Foundation early member Tyan use the POWER8... [More]

Tags:  hat openpower l4 ubuntu tyan spark controller mike opal s812lc power8 pearson habanero red powerkvm hadoop cache

    We have just had our Advanced Technology Support Early Ship Program (ESP) proto-type E850 machine refurbished at the Power Development Lab in the USA to bring it up to the General Available (GA) hardware level.  A large number of parts were replaced (these part seemed to be working fine during our test but need to get to the official level).  This is perfectly normal ESP process of removing early and test parts of a machine that is now in normal service.   What changed apart from internal electronics which... [More]

I often get asked:  "I (or my customer) thinks that the AIX system is going a bit slow!" and can I send you some nmon data. First, while nmon has performance data it is not aimed at problem diagnostics. Second, until you have a PMR (the IBM Problem Management Record) - you don't actually have a problem by definition here at IBM. Here is my boiler plate answer - I regard this as a "work in progress" as I might add further thoughts ( GREEN is a pre PMR sanity check , BLUE is the PMR preparation , RED is the... [More]

OpenStack is a fast moving open source project and rapidly adding new features. It is actually very impressive how they have so many interrelated projects (one or more for every sub-system lets call it 60) and still manage to coordinate regular testing and releases every six months.  This is of course, primarily, designed for running virtual machines in "cattle mode" = lots of them, they are all the same and if one gets sick then shot it in the head and make another one. PowerVC then takes the official newly released OpenStack... [More]

This Blog is brought to you by Clive Benjamin (Principal Engineer) who worked out how to do it. If your HMC has a quality high resolution local screen or KVM screen then the Restricted Shell terminal window font size is far too small to read and the font used very narrow lines - in our humble opinion . This blog shows you how to fix this and "no", the obvious window Font Size option does not work. 1) Below: On the HMC start the Restricted Shell window 2) Below: With the cursor on the terminal window used Control + Right mouse button... [More]

IBM Systems Director and PowerVC make use of a private API to use Shared Storage Pool (SSP) clones of LU (like a LUN allocated from the pool). But there is no command to let you do that.  LU Clones are quick and good for duplicating a master copy for small virtual machines to boot the same OS level and it means greatly reduced disk use as blocks are shared until written too. As there is not VIOS command to clone there is no command to let you see what is cloned and what is not. I tripped over this when trying to clean up my SSP which is... [More]

Most of the Power Systems range of servers since POWER7 allow for the HMC organised online hot replacement of: Cooling fans Power supplies and the online hot adding, removing or replacement of: Internal hard disks PCI(e) Adapters. From POWER8 onwards we are not going to have the online add or remove CEC nor adding or removing Processors modules nor Memory cards.  Of course online changes using Capacity Upgrade on Demand is still very much an option.  On the POWER8, I/O Drawers we don't currently have the ability online hot add or... [More]

Warning : I need to make it clear that these pictures are from an Early Ship Program machine which might vary a little from the final Generally Available (GA) models from November 2014. For example, missing labels or stickers plus colours of handles and parts etc. The engineering will be much the same.     Our machine is part of the Early Ship Program and we get to run hardware tests before machines are shipped to customers from GA.  These test include things we would never expect customer to do - which can be... [More]

We have the new HMC 8.1 code stream to manage the POWER8 machines. It also manages the POWER6 and POWER7 machines but not POWER5.  I find the HMC Virtual Storage Management panel very useful and saves me time for assigning Shared Storage Pool (SSP) space to virtual machines (VM's) and automatically doing this for Dual VIO Servers without having to log-on to the VIOS servers at all.  It is also good for assigning virtual DVDs to VMs to boot and install the various operating systems, AIX, RHEL, SLES, Fedora, Debian, etc.... [More]