AIXpert Blog is about the AIX operating system from IBM running on POWER based machines called Power Systems and software related to it like IBM Systems Director, PowerVM for virtualisation and PowerSC for security plus performance monitoring and nmon
Update: The developers and the manuals call this Load Sharing but most people think it is called Load Balancing. Perhaps, balancing gives the wrong impression of fine grain packet by packet balancing where we actually have higher level, cruder splitting of the work with Sharing. Below I use the word Balancing but mean Sharing. I have got a few questions recently on how to set this up as there are announcement with near zero information on setup, the configuration needed and a worked example. So here goes. For a long time now we have had SEA... [More]
Before we look further in to memory affinity we need to recap on the scheduling of processes and process threads of a multi-threaded process to simultaneous multi-threading (SMT) processors like POWER7. POWER5 and POWER6 had two modes of SMT off (one thread) and on (two threads) - with SMT=on two processes run at the same time (in the same clock cycle) on the CPU-core but using the different logical units inside of the CPU-core (units like the integer maths (there is more than one), floating point maths (there is more than one), compare and... [More]
I am preparing sessions for a POWER/AIX Technical Conference in Norway - unfortunately the only place they could book for the event was a Ski Resort up in the mountains and there is still good snow. It is a tough gig but someone has to do it :-) As part of that conference, I am updating them on Systems Director and demonstrating it. I had a slide with these hints and tips. A slide with zero percent marks for style (far to many words and just a long list) and I thought ... I should share these with everyone, so here they are. Most of these are... [More]
This mysterious AIX CPU Folding area is often misunderstood, so below is what I know from osmosis from talking to various guru level developers over the last 10 years. Shared Processor virtual machines (LPARs for the old fashioned) have a setting called Virtual Processors (or VP for short). This is the number of physical CPUs that the virtual machines can spread out across - in fact, I prefer to call it the "spreading factor" as it is much more obvious what it means. This can be the upper threshold for the number of CPUs that can be... [More]
The title should read "Local, Near & Far ..." - I will not correct it or links might fail. With a shared processor virtual machine (I am calling this "VM" but was called LPAR!) there are various suggestions of setting Entitlement ("Desired processing units" on the LPAR profile on the HMC, I am calling this "E") and Virtual Processor numbers (I am calling this "VP"). For Capped, the Entitlement is the maximum guaranteed CPU time that you can't go over and you round up the Entitlement to the... [More]
I have had a couple of Power systems administrators make assumptions about the virtual Ethernet speed improvements when they install a 10 Gb IVE/HEA in a VIOS which are simply not true. I guess that if three teams have made this mistake then others are about too. So I intend here to put the record straight. The expectation is that (deliberately fully spelt out long-hand to make it very clear): When they upgrade the Integrated Virtual Ethernet (also called a Host Ethernet Adapter) from 1 Gigabit per second to 10 Gigabit per second that the... [More]
So you know about Power7 Local, Near and Far memory for your actual machine but what is your Virtual Machine (LPAR) actually using? There are three key commands to show you (lssrad, mpstat and topas) and we will look at some example output. First, we need to define an SRAD or a Scheduler Resource Affinity Domain . If you have used Resource Sets with AIX WLM or WPAR then you have a good idea what these are like. An SRAD is a group of resources but in our case CPU/cores and the associated memory that is directly attached to it. As an example... [More]
Please provide comments on this blog entry - I want to know where I have it wrong?
With IBM's massive investment in Linux on Power my brain hurts trying to express the advantages. This is because the technology is excellent but the bits you want depend on where you are coming from. So I have tried to break out potential users in to various types.
For the record I am an Ab Linux on POWER User type!
A - Current AIX and POWER user
Already understands the POWER infrastructure, AIX and/or IBM i and are adding... [More]
I got asked these questions recently and had to go look the subject up ... again! I seem to have forgotten some of the details and then I thought I would use some new features of AIX for the second part. In the distant past there was various way to stop core files being dumped in to the current working directory of the program that failed. In AIX 5.3, AIX 6 and 7, the "chcore" command does all the hard work for us by letting us
Choose a specific directory for core files - which is best in it own filesystem so it... [More]
Just completed the move of the POWER and AIX Hands-on Movies to YouTube videos - after pruning we have 97 of them. This means you can watch them on your mobile Smartphone and Tablets in addition to your workstation/personal computer. You can find them all here:
YouTube look for my "Nigel Griffiths" Channel http://www.youtube.com/user/nigelargriffiths This should mean they are simpler to find, search for and view. During the transfer I pruned out a few of the older ones particular Systems 6.2 as we now have a set of Systems... [More]
I thought I should summarise the long eleven part Local, Near & Far POWER7 Affinity series. 1) Placement: Find out the layout of your boxes CPU and RAM and if the RAM is evenly distributed across available DIMMs Find out the placement of your Virtual Machines (LPARs) with lssrad -av - or - topas -M 2) SMT4 : Expect POWER7 SMT4 CPU use to “look” different POWER5 & 6 have two equal threads POWER7 shuts down threads 3 & 4 and even thread 2 - when there is not enough processes running. 3) Entitlement : Only set minimum Entitlements,... [More]
I just realised that I got help from the AIX developers to explain the output of the mpstat -d command for my POWER7 and Affinity Technical University sessions in Miami and Copenhagen that I never passed on to every one. In an earlier blog we examined some mpstat columns but skipped others. The nearly undocumented mpstat stats are: S0rd, S1rd, S2rd, S3rd, S4rdand S5rd. In the AIX manuals it states: S0rd = The percentage of thread re-dispatches
within the scheduling affinity domain 0.
And likewise for the other numbers - but what does... [More]
I have been wondering why the lower end POWER7 machines have local and far memory and not local and near. Perhaps you wondered too! Well at the Miami Power Technical University, I got to talk to Dr Joel Tendler (IBMer) and a POWER7 processor guru and put the question to him. He covered this sort of architectural topic in his presentation at the event and I learnt a lot in this area by listening to the "master". Below is some background and the explanation too. The POWER7 chips has two memory controllers for maximum performance but... [More]
N-Port Id Virtualisation (NPIV) and virtual Fibre Channel adapters - like high speed physical adapters (see an earlier blog) do require memory to operator at full speed. NPIV effectively turns the Virtual I/O Server (VIOS) into a kind of virtual SAN switch where packets are passed through with no changes from the physical adapter to the client Virtual Machine (LPAR). I found it very hard to find a recommendation in the documentation but was given a rule of thumb. As before it should be noted that starving the VIOS and the Hypervisor of... [More]
I was just asked this question by customer and it got me thinking, Items I would include: WPAR takes seconds to create and LPARs minutes LPARs need setting VIOS LV or LUNs but WPARs add simpler NFS mount points, or can just use Global AIX diskspace options LPAR needs 512 to 1GB to boot AIX and a WPAR takes just ~60 MB (yes sixty megabytes) You can share application code say 1 GB in each and every LPAR (40 LPAR = 40 GB) or just one shared read-only copy for all WPARs 40 WPAR = 1 GB).This saves man-power in maintenance, disk space AND memory (if... [More]