It seems I am not alone on this annoying RHEL 7 missing feature so I thought I would share the solution.   Note: Installing Python 3 on Ubuntu 16 or 18 takes about 3 seconds. Note: Only historic applications use Python 2.7 and around the web, you will find lots of comments about ONLY write Python 3.   So I need Python 3 on RHEL - I thought this would take only a few seconds using yum on Red Hat as I have the repositories already setup. But installing Python 3 on REHL 7.6 - it took me about 2 hours of frustration. Python 2.7 seems to be there at... [More]

I have a small virtual machine with two disks that I want information about my disks, so I look in: /proc/diskstats the file about disks right? I find: $ cat /proc/diskstats 8 0 sda 157237 6081 20263539 1230080 479407 147344 44907178 13965510 0 763580 15201720 8 1 sda1 51 0 8832 30 0 0 0 0 0 20 30 8 2 sda2 53 0 11136 50 0 0 0 0 0 40 50 8 3 sda3 157105 6081 20236531 1229970 479407 147344 44907178 13965510 0 763540 15201610 8 16 sdb 1829 0 116291 1370 102 3 31551 59640 0 2240 61010 8 17... [More]

#LinuxNeedsFIXING part 1: On my largely idle POWER8 server with 20 CPU cores (160 CPU core threads), when I run the regular command:       ps -ef it lists 1317 processes = 1317 lines of output!! This results in 35 screens full of pointless process stats zooming off the top of my screen. That is 86KB dumped to my screen and 99% of it never gets read - I can't even scroll back to the start. The problem is 1270 lines are kernel processes 970 with zero seconds of CPU time and the rest less than 10 CPU seconds in a year of server uptime! The... [More]

I was recently invited to present in Oslo, Norway on this topic to not a very technical audience. Which was outside my regular comfort zone. Other speakers were covering IBM i - the world's most modern platform HPC from Experience (Norway's universities are amongst world leaders in this area) and a client speaking on their experience migrating to Oracle to POWER9 (the S922 servers were faster than expected and they are now purchasing E950s).   My topic was the same as this Blog title but I added a few comments on AIX. Here is what I... [More]

I have had a number of clients get into a mess with an AIX performance CritSit (Critical Situation) but they have forgotten to capture nmon data before the incident. So there is no "before and after "comparisons that can be made to highlight the change causing issues and likely causes.   Well, you might not know but AIX is automatically capturing performance data into the directory /etc/perf/daily/  by default. No action required by you. Now the bad news - it is using topasrec and recording it as raw binary data. Now the good news you... [More]

To get protection from the Spectre / Meltdown security issues you need a few items in place: 1) A systems firmware level that supports the protection All POWER9 systems firmware has this builtin You may need to upgrade your POWER7 or  POWER8 firmware to a recent version - which is Best Practice anyway 2)  The system firmware protection is actually switched on To check this use the HMC -> ASMI -> "System Configuration“ ->  "Speculative Execution Control“ To change the setting, first Power-Off the server (sorry) then go to the... [More]

I am getting this reported regularly. The techie usually confirms the server is in Max-Performance mode - Which would be my first question. Typically, don't say how they are monitoring the GHz. I am 99% sure they are using older commands and thus listing the nominal frequency and than never changes and mot the current frequency which bounces up to the highest GHz and stays there.  If the POWER9 is very large config = maximum CPU cores + maximum memory size + loads of high speed adapter (that can generate heat too) + if the model has... [More]

So I took the latest SUSE SLES 15 DVD for POWER to install them on my latest POWER9 server (S924) running the PowerVM Hypervisor and 100% failed to get this installed .  I asked around and no one was much help!   Below is what I did not know or realise at the time that cause my failure - in my defence they are not clearly documentation or assumed.   As an IBMer I can run SUSE SLES (and Red Hat RHEL) for skills enhancement and demos without paying for the licence. This is an excellent deal but then I would be demonstrating SLES to customers and... [More]

I am not going to cover actual Python coding here (well, may be, a little at the end) but the good and bad places to start and things to avoid. First used at the Hollywood, Florida and Rome, Italy IBM Technical University conferences - we call them TechU   Alternatives You could just search Google, YouTube, and many other places and find 10 billion hits You will quickly get totally swamped with options This is Nigel's starter pack for  a quick start. This is what I found very useful - You, of course, may be different !!! What... [More]

In 2016, I setup a Raspberry Pi with five temperature probes to measure in my computer room: Air Conditioning output temperature Room Temperature Three Temperatures at the rear exhaust of three larger servers You can find the article here:  Computer_Room_Temperature_Monitoring_with_a_Raspberry_Pi   I end up with a script that gets the date+time and 5 temperatures.  The data is saved in a Comma Separated Values file and then I run a bunch of scripts to generate .html web pages using Google Chart JavaScript graphs.  It took a... [More]

I get this question about twice a month so I thought I would answer it here and refer to this Blog Entry.   This is a non-trivial question as it is complicated.   If you are a " just give me the basics on the back of this envelope " sort of person then do one of the three below: If you want the real easy answer hire me for 10,000 €$£ (Euro Dollar Pounds) a day plus expenses and I will do it for you and teach you at the same time via "1 to 1" skills transfer. Or phone your local IBM Representative and get... [More]

A couple of times a month I get a question like this:  I have been sent 35 GB of nmon files covering the last month, about 6 servers with 40 LPARs (virtual machines) each server in total ~6000 files. How do I load that in to the nmon Analyser to see the graphs? My first reaction is: that is not my problem I wrote nmon for AIX and nmon for Linux and not the nmon analyser. If you stop and think, it is obvious that you are never ever going to load 35 GB in to Excel on a laptop with 8 to 16 GB of memory. Excel is likely to grind the laptop a... [More]

If you don't know VIOS is short for Virtual I/O Server then you are probably on the wrong page :-) NEW NEWS - May 2019 Updated VIOS levels became available in April/May 2019 Latest VIOS 2.2.6 version and recommended as the start point to upgrade is VIOS 2.2.6.41 (was 2.2.6.32) Getting to this level before the upgrade will remove possible problems. Latest VIOS 3.1 version and recommended as your final target is VIOS 3.1.0.21 (was 3.1.0.10) This has 270 fixes that are best added before you return your VIOS 3.1 to production mode.  ... [More]

I have been testing out the AIX libperfstat libraries new functions and got to the SSP ones. So I knock up this command to dump the config details that can be extracted.   So with 128 lines of C code we have the new command - I have not seen a new VIOS SSP command that outputs this data yet so this is new details that I think would help you rebuild a SSP if you had an accident or disaster. Command line interface # ./nsspconf -h Hint: [-v] [-h] List Shared Storage Pool info: global, disk including tier & failure group, LU... [More]