Category Archives: Virtualization

Swap, swappiness, Servers not responding

I have read a lot of wrong recommendations about the use of Swap and Swappiness so I want to bring some light about it.

The first to say is that every project is different, so it is not possible to make a general rule. However in most of the cases we want systems to operate as fast and efficiently as possible.

So this suggestions try to covert 99% of the cases.

By default Linux will try to be as efficient as possible. So for example, it will use Free Memory to keep IO efficient by keeping in Memory cache and buffers.

That means that if you are using files often, Linux will keep that information cached in RAM.

The swappiness Kernel setting defines what tradeoff will take Linux between keeping buffers with Free Memory and using the available Swap Memory.

# sysctl vm.swappiness
vm.swappiness = 60

The default value is 60 and more or less means that when RAM memory gets to 60%, swap will start to be used.

And so we can find Servers with 256GB of RAM, that when they start to use more than 153 GB of RAM, they start to swap.

Let’s analyze the output of free -h:

carles@vbi78g:~/Desktop/Software/checkswap$ free -h
              total        used        free      shared  buff/cache   available
Mem:          2.9Gi       1.6Gi       148Mi        77Mi       1.2Gi       1.1Gi
Swap:         2.0Gi        27Mi       2.0Gi

So from this VM that has 2.9GB of RAM Memory, 1.6GB are used by applications.

The are 148MB that can immediately used by Applications, and there are 1.2GB in buffers/cache. Does that means that we can only use 148MB (plus swap)?. No, that mean that Linux tried to optimize io speed by keeping 1.2GB of RAM memory in buffers. But this is the best effort of Linux to have performance, for real applications will be also able to use 1.1GB that corresponds to the available field.

About swap, from 2GB, only 27MB have been used.

As vm.swappiness is set to 60, more RAM will be swapped out to swap, even if we have lots available.

As I said every case is different. If we are talking about a Desktop that has NVMe drives, the impact will be low. But if we are talking about a Server that is a hypervisor running VMs and has high usage on CPU and has the swap partition or the swap in a file, that could lead to huge problems. If there is a physical Server with a single spinning drive (or logical unit through RAID), and one partition is for Swap, and the other for mountpoints, and a process is heavily reading/writing to a partition mounted (an elastic search, or a telegraf, prometheus…), and the System tries to swap, then they will be competing for the magnetic head of disk, slowing down everything.

If you take a look on how the process of swapping memory pages from the memory to disk, you will understand that applications may need certain pages before being able to run, so in many cases we get to lock situations, that force everything to wait.

In my career I found Servers that temporarily stopped responding to ping. After a while ping came back, I was able to ssh and uptime showed that the Server did not reboot.

I troubleshooted that, and I saw a combination of high CPU usage spikes and Swap usage.

Using iostat and iotop I monitored what was speed of transference of only 1 MB/second!!.

I even did swapoff and it took one hour to free 4 GB swap partition!.

I also saw swap partition being in a spinning disk, and in another partition of the same spinning drive, having a swapfile. Magnetic spinning drives can only access one are of the drive at the same time, so that situation, using swap is very bad.

And I have seen situations were the swap or swapfile was mounted in a block device shared via network with the Server (like iSCSI or NFS), causing terrible performance when swapping.

So you have to adapt the strategy according to the project.

My preferred strategy for Compute Nodes and NoSQL Databases is to not use swap at all. In other cases, like MySQL Databases I may set swappiness to preferably to 1 or to 10.

I quote here the recommendations from couchbase docs:

The Linux kernel’s swappiness setting defines how aggressively the kernel will swap memory pages versus dropping pages from the page cache. A higher value increases swap aggressiveness, while a lower value tells the kernel to swap as little as possible to disk and favor RAM. The swappiness range is from 0 to 100, and most Linux distributions have swappiness set to 60 by default.

Couchbase Server is optimized with its managed cache to use RAM, and is capable of managing what should be in RAM and what shouldn’t be. Allowing the OS to have too much control over what memory pages are in RAM is likely to lower Couchbase Server’s performance. Therefore, it’s recommended that swappiness be set to the levels listed below.

https://docs.couchbase.com/server/current/install/install-swap-space.html

Another theme, is when you log to a Server and you see all the Swap memory in use.

Linux may have moved the pages that were less used, and that may be Ok for some cases, for example a Cron Service that waits and runs every 24 hours. It is safe to swap that (as long as the swap IO is decent).

When Kernel Swaps it may generate locks.

But if we log to a Server and all the Swap is in use, how can we know that the Swap has been quiet there?.

Well, you can use iostat or iotop or you can:

cat /proc/vmstat

This file contains a lot of values related to Memory, we will focus on:

pswpin 508992338
pswpout 280871088

In https://superuser.com/questions/785447/what-is-the-exact-difference-between-the-parameters-pgpgin-pswpin-and-pswpou you can find very interesting description of those values. I paste here an excerpt:

Paging refers to writing portions, termed pages, of a process’ memory to disk.
Swapping, strictly speaking, refers to writing the entire process, not just part, to disk.
In Linux, true swapping is exceedingly rare, but the terms paging and swapping often are used interchangeably.

page-out: The system’s free memory is less than a threshold “lotsfree” and unnused / least used pages are moved to the swap area.
page-in: One process which is running requested for a page that is not in the current memory (page-fault), it’s pages are being brought back to memory.
swap-out: System is thrashing and has deactivated a process and it’s memory pages are moved into the swap area.
swap-in: A deactivated process is back to work and it’s pages are being brought into the memory.

Values from /proc/vmstat:

pgpgin, pgpgout – number of pages that are read from disk and written to memory, you usually don’t need to care that much about these numbers

pswpin, pswpout – you may want to track these numbers per time (via some monitoring like prometheus), if there are spikes it means system is heavily swapping and you have a problem.

In this actual example that means that since the start of the Server there has been 508992338 Page Swap In (with 4K memory pages this is 1,941 GB, so almost 2 TB transferred) and for Page Swat Out (with 4K memory pages this is 1,071 GB, so 1 TB of transferred). I’m talking about a Server that had a 4GB swap partition in a spinning disk and a 12 GB swapfile in another ext4 partition of the same spinning disk.

The 16 GB of swap were in use and iotop showed only two sources of IO, one being 2 VMs writing, another was a journaling process writing to the mountpoint where the swapfile was. That was an spinning drive (underlying hardware was raid, for simplicity I refer to one single drive. I checked that both spinning drives were healthy and fast). I saw small variations in the size of the Swap, so I decided to monitor the changes in pswpin and pswpout in /proc/vmstat to see how much was transferred from/to swap.

I saw then how many pages were being transferred!.

I wrote a small Python program to track those changes:

https://gitlab.com/carles.mateo/checkswap

This little program works in Python 2 and Python 3, and will show the evolution of pswpin and pswpout in /proc/vmstat and will offer the average for last 5 minutes and keep the max value detected as well.

As those values show the page swaps since the start of the Server, my little program, makes the adjustments to show the Page Swaps per second.

A cheap way to reproduce collapse by using swap is using VirtualBox: install an Ubuntu 20.04 LTS in there, with 2 GB of less of memory, and one single core. Ping that VM from elsewhere.

Then you may run a little program like this in order to force it to swap:

#!/usr/bin/env python3
a_items = []
i_total = 0
# Add zeros if your VM has more memory
for i in range(0, 10000000):
    i_total = i_total + i
    a_items.append(i_total)

And checkswap will show you the spikes:

Many voices are discordant. Some say swappiness default value of 60 is good, as Linux will use the RAM memory to optimize the IO. In my experience, I’ve seen Hypervisors Servers running Virtual Machines that fit on the available physical RAM and were doing pure CPU calculations, no IO, and the Hypervisor was swapping just because it had swappiness to 60. Also having swap on spinning drives, mixing swap partition and swapfile, and that slowing down everything. In a case like that it would be much better not using Swap at all.

In most cases the price of Swapping to disk is much more higher than the advantage than a buffer for IO brings. And in the case of a swapfile, well, it’s also a file, so my suspect is that the swapfile is also buffered. Nothing I recommend, honestly.

My program https://gitlab.com/carles.mateo/checkswap may help you to demonstrate how much damage the swapping is doing in terms of IO. Combine it with iostat and iotop --only to see how much bandwidth is wasted writing and reading from/to swap.

You may run checkswap from a screen session and launch it with tee so results are logged. For example:

python3 checkswap.py | tee 2021-05-27-2107-checkswap.log

If you want to automatically add the datetime you can use:

python3 checkswap.py | tee `date +%Y-%m-%d-%H%M`-checkswap.log

Press CTRL + a and then d, in order to leave the screen session and return to regular Bash.

Type screen -r to resume your session if this was the only screen session running in background.

An interesting reflection from help Ubuntu:

The “diminishing returns” means that if you need more swap space than twice your RAM size, you’d better add more RAM as Hard Disk Drive (HDD) access is about 10³ slower then RAM access, so something that would take 1 second, suddenly takes more then 15 minutes! And still more then a minute on a fast Solid State Drive (SSD)…

https://help.ubuntu.com/community/SwapFaq

Do you have a swap history that you want to share?.

Cloning a Windows Application running in Wine

I’ve some very old Windows Applications running in Wine in my Linux workstations.

It’s Software I bought years ago and that is not available anymore.

Keeping and migrating or cloning to another Linux Workstation or Virtual machine is really easy.

I share the steps with you.

You just have to copy the contents from your /home/username/.wine folder.

Then, in the new workstation install wine. For Ubuntu this is:

sudo apt update && sudo apt install wine

Run winecfg so basic links and structures are created.

Then simply copy the .wine folder backup to your new machine /home/username/

Your programs will be in /home/username/.wine/drive_c_/Program Files/ or /home/username/.wine/drive_c_/Program Files (x86)/

If you want you can just copy your programs folder.

Remember that to cd to a directory with spaces you have to use “

For example:

$ pwd
/home/carles/.wine/drive_c
$ cd "Program Files"
$ pwd
/home/carles/.wine/drive_c/Program Files

You can also use \ (slash space) to escape space.

Then start your favorite program with:

wine yourprogram.exe

If that fails is very probably that creating a new configuration, for a new user, will make things right.

Update 2022-01-05: Take in count that you will be copying the Windows registry when doing this. I use this trick to clone applications that are no longer downloadable from the Internet. I clone wine to dedicated Virtual Machines. You may need different Virtual Machines for different programs if windows registry is different for them.

A base Dockerfile for my Jenkins deployments

Update: I’ve created a video and article about how to install jenkins in Docker with docker CLI and Blue Ocean plugins following the official Documentation. You may prefer to follow that one.

Update: Second part of this article: Creating Jenkins configurations for your projects

So I share with you my base Jenkins Dockerfile, so you can spawn a new Jenkins for your projects.

The Dockerfile installs Ubuntu 20.04 LTS as base image and add the required packages to run jenkins but also Development and Testing tools to use inside the Container to run Unit Testing on your code, for example. So you don’t need external Servers, for instance.

You will need 3 files:

  • Dockerfile
  • docker_run_jenkins.sh
  • requirements.txt

The requirements.txt file contains your PIP3 dependencies. In my case I only have pytest version 4.6.9 which is the default installed with Ubuntu 20.04, however, this way, I enforce that this and not any posterior version will be installed.

File requirements.txt:

pytest==4.6.9

The file docker_run_jenkins.txt start Jenkins when the Container is run and it will wait until the initial Admin password is generated and then it will display it.

File docker_run_jenkins.sh:

#!/bin/bash

echo "Starting Jenkins..."

service jenkins start

echo "Configure jenkins in http://127.0.0.1:8080"

s_JENKINS_PASSWORD_FILE="/var/lib/jenkins/secrets/initialAdminPassword"

i_PASSWORD_PRINTED=0

while [ true ];
do
    sleep 1
    if [ $i_PASSWORD_PRINTED -eq 1 ];
    then
        # We are nice with multitasking
        sleep 60
        continue
    fi

    if [ ! -f "$s_JENKINS_PASSWORD_FILE" ];
    then
        echo "File $s_FILE_ORIGIN does not exist"
    else
        echo "Password for Admin is:"
        cat $s_JENKINS_PASSWORD_FILE
        i_PASSWORD_PRINTED=1
    fi
done

That file has the objective to show you the default admin password, but you don’t need to do that, you can just start a shell into the Container and check manually by yourself.

However I added it to make it easier for you.

And finally you have the Dockerfile:

FROM ubuntu:20.04

LABEL Author="Carles Mateo" \
      Email="jenkins@carlesmateo.com" \
      MAINTAINER="Carles Mateo"

# Build this file with:
# sudo docker build -f Dockerfile -t jenkins:base .
# Run detached:
# sudo docker run --name jenkins_base -d -p 8080:8080 jenkins:base
# Run seeing the password:
# sudo docker run --name jenkins_base -p 8080:8080 -i -t jenkins:base
# After you CTRL + C you will continue with:
# sudo docker start
# To debug:
# sudo docker run --name jenkins_base -p 8080:8080 -i -t jenkins:base /bin/bash

ARG DEBIAN_FRONTEND=noninteractive

ENV SERVICE jenkins

RUN set -ex

RUN echo "Creating directories and copying code" \
    && mkdir -p /opt/${SERVICE}

COPY requirements.txt \
    docker_run_jenkins.sh \
    /opt/${SERVICE}/

# Java with Ubuntu 20.04 LST is 11, which is compatible with Jenkins.
RUN apt update \
    && apt install -y default-jdk \
    && apt install -y wget curl gnupg2 \
    && apt install -y git \
    && apt install -y python3 python3.8-venv python3-pip \
    && apt install -y python3-dev libsasl2-dev libldap2-dev libssl-dev \
    && apt install -y python3-venv \
    && apt install -y python3-pytest \
    && apt install -y sshpass \
    && wget -qO - https://pkg.jenkins.io/debian-stable/jenkins.io.key | apt-key add - \
    && echo "deb http://pkg.jenkins.io/debian-stable binary/" > /etc/apt/sources.list.d/jenkins.list \
    && apt update \
    && apt -y install jenkins \
    && apt-get clean

RUN echo "Setting work directory and listening port"
WORKDIR /opt/${SERVICE}

RUN chmod +x docker_run_jenkins.sh

RUN pip3 install --upgrade pip \
    && pip3 install -r requirements.txt


EXPOSE 8080


ENTRYPOINT ["./docker_run_jenkins.sh"]

Build the Container

docker build -f Dockerfile -t jenkins:base .

Run the Container displaying the password

sudo docker run --name jenkins_base -p 8080:8080 -i -t jenkins:base

You need this password for starting the configuration process through the web.

Visit http://127.0.0.1:8080 to configure Jenkins.

Configure as usual

Resuming after CTRL + C

After you configured it, on the terminal, press CTRL + C.

And continue, detached, by running:

sudo docker start jenkins_base

The image is 1.2GB in size, and will allow you to run Python3, Virtual Environments, Unit Testing with pytest and has Java 11 (not all versions of Java are compatible with Jenkins), use sshpass to access other Servers via SSH with Username and Password…

Solving the problem when running a Docker Container: standard_init_linux.go:190: exec user process caused “no such file or directory”

When you see this error for the first time it can be pretty ugly to detect why it happens.

At personal level I use only Linux for my computers, with an exception of a windows laptop that I keep for specific tasks. But my employers often provide me laptops with windows.

I suffered this error for first time when I inherited a project, in a company I joined time ago. And I suffered some time later, by the same reason, so I decided to explain it easily.

In the project I inherited the build process was broken, so I had to fix it, and when this was done I got the mentioned error when trying to run the Container:

standard_init_linux.go:190: exec user process caused "no such file or directory"

The Dockerfile was something like this:

FROM docker-io.battle.net/alpine:3.10.0

LABEL Author="Carles Mateo" \
      Email="docker@carlesmateo.com" \
      MAINTAINER="Carles Mateo"

ENV SERVICE cservice

RUN set -ex

RUN echo "Creating directories and copying code" \
    && mkdir -p /opt/${SERVICE}
    
COPY config.prod \
    config.dev \
    config.st \
    requirements.txt \
    utils.py \
    cservice.py \
    tests/test_cservice.py \
    run_cservice.sh \
    /opt/${SERVICE}/

RUN echo "Setting work directory and listening port"
WORKDIR /opt/${SERVICE}
EXPOSE 7000

RUN echo "Installing dependencies" \
    && apk add build-base openldap-dev python3-dev py-pip \
    && pip3 install --upgrade pip \
    && pip3 install -r requirements.txt \
    && pip3 install pytest

ENTRYPOINT ["./run_cservice.sh"]

So the project was executing a Bash script run_cservice.sh, via Dockerfile ENTRYPOINT.

That script would do the necessary amends depending if the Container is launched with prod, dev, or staging parameter.

I debugged until I saw that the Container never executed this in the expected way.

A echo “Debug” on top of the Bash Script would be enough to know that very basic call was never executed. The error was first.

After much troubleshooting the Container I found that the problem was that the Bash script, that was copied to the container with COPY in the Dockerfile, from a Windows Machines, contained CRLF Windows carriage return. While for Linux and Mac OS X carriage return is just a character, LF.

In that company we all use Windows. And trying to build the Container worked after removing the CRLF, but the Bash script with CRLF was causing that problem.

When I replace the CRLF by Unix type LF, and rebuild the image, and ran the container, it worked lovely.

A very easy, manual way, to do this in Windows, is opening your file with Notepad++ and setting LF as carriage return. Save the file, rebuild, and you’ll see your Container working.

Please note that in the Dockerfile provided I install pytest Framework and a file calles tests/test_cservice.py. That was not in the original Dockerfile, but I wanted to share with you that I provide Unit Testing that can be ran from a Linux Container, for all my projects. What normally I do is to have two Dockerfiles. One for the Production version to be deployed, another for running Unit Testing, and some time functional testing as well, from inside the Docker Container. So strictly speaking for the production version, I would not copy the tests/test_cservice.py and install pytest. A different question are internal Automation Tools, where it may be interested providing a All-in-One image, that can run the Unit Testing before start the service. It is interesting to provide some debugging tools in out Internal Automation Tools, so we can troubleshoot what’s going on in case of problems. Take a look at my previous article about Python version for Docker and Automation tools, for more considerations.

Extend existing Single ZFS disk with a mirror without losing the Data on the existing HDD

This is an answer that I did to a question in askubuntu.

https://askubuntu.com/questions/1301828/extend-existing-single-disk-zfs-with-a-mirror-without-formating-the-existing-hdd/

Question:

I have one HDD formatted as single disc zfs system on my server. It looks like the following:

Now I want to convert this to a zfs mirror without formatting the original disk. Any ideas?

Result should be something like:

hdd0
   mirror0
       ata-........................
       ata-........................

Answer:

I reproduced your case in a VM and paste here step by step. :)

Note: First of all, please do a backup of your data. I added an empty new disk, so ZFS had no doubt what was the master drive. Although you should have no problem as the first drive already forms part of the pool, a backup is recommended.

Quick answer: You need the zpool attach command.

Basically:

sudo zpool attach hdd0 existinghdd blankhdd

After, do:

zpool status

And you will see that a mirror has been created. Your data on the already existing drive will be keep, and will be replicated to the new one (Resilvered).

As ZFS only copys the actual information this process will take more or less depending on the amount of Data.

In my VM 300 GB were replicated in 3 seconds, while my experience with SAS and SATA drives, I was Resilvering 10 TB in less than 24 hours (for that I was using drives from 10TB to 14TB SAS) .

Now the long answer with everything I did in my Virtual Box VM:

lsblk --scsi

identify the two empty drives by:

ls /dev/disk/by-id/

Select one of them and create a pool like your: sudo zpool create hdd0 id_of_mydrive

See that pool /hdd0 has been created and mounted on root.

sudo zpool status sudo zpool list sudo ls -al /hdd0

Fill with some random data (or better copy files there) to generate a drive like data like you. I generated from random:

sudo dd id=/dev/urandom of=/hdd0/file.000 bs=1M count=100 status=progress
sudo dd id=/dev/urandom of=/hdd0/file.001 bs=1M count=100 status=progress
sudo dd id=/dev/urandom of=/hdd0/file.002 bs=1M count=100 status=progress

Then I got the checksum and saved to verify later.

sudo su
# Please note I continue as root
sha512sum file.000 > file.000.sha512
sha512sum file.001 > file.001.sha512
sha512sum file.002 > file.002.sha512

zpool list shows nearly 100GB of space.

zpool attach hdd0 id_of_mydrive id_of_the_drive_to_add

zpool status will show:

pool: hdd0
state: ONLINE
scan: resilvered 301M in 0 days 00:00:03 with 0 errors…

   NAME                            STATE   READ WRITE CKSUM   
   hdd0
     mirror-0
       ata-VBOX_HARDDISK_VBa8...   ONLINE     0     0     0
       ata-VBOX_HARDDISK_VB8c...   ONLINE     0     0     0

errors: No known data errors

I verified the checksums.

zpool list will return as well 99GB of space available, as two drives of 100GB are being used in mirror.

So as kaulex mentioned the format is: zpool attach

Where device is your previous vdev with data (the single hard drive with Data in the ZFS pool named ‘hdd0’).

As I did you want to use the Id of the device and not the name, so you will use the identifier in /dev/disk/by-id/ and not sdb, sdc… (Please note, adding /dev/ is not necessary). The reason to do not use device names like sdb, sdc, sdea, etc… is that those names may change why live is running or between reboots. The id never changes. In real systems, not Virtual Box, they may start by wwn or ata.

Install Windows Subsystem for Linux, WSL 2 on Windows 10 64 bit, with Ubuntu, solution to error WslRegisterDistribution failed with error: 0x80070057

You know I love Linux. I was compiling my own Kernels back in 1995, when it took more than 24 hours in a 386, and working on the first ISPs in Barcelona managing the Linux Systems.

For my computers I prefer Linux, no doubt about it, but many multinationals I worked for have Windows option only for the Laptops and Desktops.

During years I had to deal with sending files to Linux or Unix (HP UX, Sun Solaris…) to process them and getting back the result. Some sort of ETL and Map Reduce in the prehistory of personal computers, taking in count aspects like Networks speeds too, available space, splitting files for processing.

When I was working as Senior Project Manager in Winterthur Insurance, now Axa, I had to run a lot of ETL (Extract Transform Load) for considerably big files, or when I was project manager and later head of department in Volkswagen gedas or later helping Start ups like Privalia. I can tell you that Windows didn’t like you to open editors to work with 1GB text or CSV file, and doesn’t like it, even if your computer has 16GB of Memory, and even if they do the simplicity of Bash scripts and using pipes, grep, awk… is so powerful that is very convenient to have those files processed using Linux.

And honestly is a pain to send back and forth files to a UNIX System just for Data Crunch. And a VM will be slow and use memory, and you have enable some sort of sharing with it so it can access the Data. Not to talk if you need to split the data files in blocks to be processed in parallel by several computers.

There are many solutions, like using Virtual Machines, Docker, external Servers, etc…

WSL allows you to run Linux command line tools inside Windows.

https://wiki.ubuntu.com/WSL

Having WSL allows things to be done much more straightforward, processing the files in your local windows hard drives.

Please note: Maybe you have enough using GitBash.

Error installing: WslRegisterDistribution failed with error: 0x80070057

When I installed it I found this error and look for an answer online. I found no solutions and many people suffering from the same problem, so I decided to publish an article on how to make it work.

The instructions I detail are based on the documentation from Canonical https://wiki.ubuntu.com/WSL and from Microsoft https://docs.microsoft.com/en-us/windows/wsl/install-win10 and the final solution I found.

Microsoft use Powershell to activate the features disabled in Windows, I did the same with Command Line, which I found more convenient for most of the non extremely tech people.

You will need:

  • For x64 systems: Version 1903 or higher, with Build 18362 or higher.

You can check your version of windows opening a Terminal (CMD.exe) and typing:

winver
  • For ARM64 systems: Version 2004 or higher, with Build 19041 or higher.

I’m not covering installing WSL for ARM, only for Intel/AMD Desktop/Laptops with Windows 10.

If you’re unsure, you can open a Terminal (CMD.exe) and run:

systeminfo | find "System Type"

If is a x64 system it will return: x64-based PC

Launch CMD.exe as Administrator and type:

dism /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart

Enable Virtual Machine Feature:

dism /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart

At this point you have to Restart Windows to complete this part of the installation, otherwise next step won’t work.

Download the Linux Kernel Update Package from here:

https://wslstorestorage.blob.core.windows.net/wslblob/wsl_update_x64.msi

Execute wsl_update_x64.msi and grant permissions for modifying the system.

Now it’s crucial that you reboot again. Even if you’re not asked for.

That’s the tricky part.

Then set the version 2 of WSL as default:

C:\WINDOWS\system32>wsl --set-default-version 2
 For information on key differences with WSL 2 please visit https://aka.ms/wsl2

Installing Ubuntu (or Kali, or Debian, or openSUSE…)

Open the Microsoft Store or use Microsoft Store link to go directly to the installation of Ubuntu.

You have also several options as indicated in https://wiki.ubuntu.com/WSL:

The recommended way to install Ubuntu on WSL is through the Microsoft Store.

The following Ubuntu releases are available as apps on the Microsoft Store:

  • Ubuntu 16.04 LTS (Xenial) is the first release available for WSL. It supports the x64 architecture only. (offline installer: x64)
  • Ubuntu 18.04 LTS (Bionic) is the second LTS release and the first one supporting ARM64 systems, too. (offline installers: x64, ARM64)
  • Ubuntu 20.04 LTS (Focal) is the current LTS release, supporting both x64 and ARM64 architecture.
  • Ubuntu (without the release version) always follows the recommended release, switching over to the next one when it gets the first point release. Right now it installs Ubuntu 20.04 LTS.

Each app creates a separate root file system in which Ubuntu shells are opened but app updates don’t change the root file system afterwards. Installing a different app in parallel creates a different root file system allowing you to have both Ubuntu LTS releases installed and running in case you need it for keeping compatibility with other external systems. You can also upgrade your Ubuntu 16.04 to 18.04 by running ‘do-release-upgrade’ and have three different systems running in parallel, separating production and sandboxes for experiments.

But if you prefer, instead of using the Windows Store, you can download the appx.

In the same page mentioned you can do it for several versions, I attach the link for Ubuntu 20.04 LTS: https://aka.ms/wslubuntu2004

Assuming you used the Windows Store, if you did not reboot and try now to execute it for the first time, or you go to the Command Line and write bash, or open Ubuntu from Windows menu, whatever method you use, you’ll get the abovementioned error.

If that happens to you, just reboot and when you open it will work and will start the install and ask for a user and password:

From here you’re able to update the system, execute the text commands available in Linux, access to the Windows drives, launch htop, git, Python3, apt, wget… copy and paste between windows and Linux terminal, share PATH…

And of course you can run CTOP.py

Take in count that the space reported in / partition is not real, and that you have a 4GB swap.

You can access your C:\ Windows files from:

/mnt/c/

Refreshing settings in a Docker immutable image with Python and Flask

This is a trick to restart a Service that is running on a immutable Docker, with some change, and you need to refresh the values very quickly without having to roll the CI/CD Jenkins Pipeline and uploading a new image.

So why would you need to do that?.

I can think about possible scenarios like:

  • Need to roll out an urgent fix in a time critical manner
  • Jenkins is broken
  • Somebody screw it on the git master branch
  • Docker Hub is down
  • GitHub is down
  • Your artifactory is down
  • The lines between your jumpbox or workstation and the secure Server are down and you have really few bandwidth
  • You have to fix something critical and you only have a phone with you and SSH only
  • Maybe the Dockerfile had latest, and the latest image has changed
FROM os:latest

The ideal is that if you work with immutable images, you roll out a new immutable image and that’s it.

But if for whatever reason you need to update this super fast, this trick may become really handy.

Let’s go for it!.

Normally you’ll start your container with a command similar to this:

docker run -d --rm -p 5000:5000 api_carlesmateo_com:v7 prod 

The first thing we have to do is to stop the container.

So:

docker ps

Locate your container across the list of running containers and stop it, and then restart without the –rm:

docker stop container_name
docker run -d -p 5000:5000 api_carlesmateo_com:v7 prod

the –rm makes the container to cleanup. By default a container’s file system persists even after the container exits. So don’t start it with –rm.

Ok, so login to the container:

docker exec -it container_name /bin/sh 

Edit the config you require to change, for example config.yml

If what you have to update is a password, and is encoded in base64, encode it:

echo -n "ThePassword" | base64
VGhlUGFzc3dvcmQ=

Stop the container. You can do it by stopping the container with docker stop or from inside the container, killing the listening process, probably a Python Flask.

If your Dockerfile ends with something like:

ENTRYPOINT ["./webservice.py"]

And webservice.py has Python Flask code similar to this:

#!/usr/bin/python3
#
# webservice.py
#
# Author: Carles Mateo
# Creation Date: 2020-05-10 20:50 GMT+1
# Description: A simple Flask Web Application
#              Part of the samples of https://leanpub.com/pythoncombatguide
#              More source code for the book at https://gitlab.com/carles.mateo/python_combat_guide
#


from flask import Flask, request
import logging

# Initialize Flask
app = Flask(__name__)


# Sample route so http://127.0.0.1/carles
@app.route('/carles', methods=['GET'])
def carles():
    logging.critical("A connection was established")
    return "200"

logging.info("Initialized...")

if __name__ == "__main__":
    app.run(host='0.0.0.0', port=5000, debug=True)

Then you can kill the process, and so ending the container, from inside the container by doing:

ps -ax | grep webservice
 5750 root     56:31 {webservice.py} /usr/bin/python /opt/webservice/webservice.py
kill -9 5790

This will finish the container the same way as docker stop container_name.

Then start the container (not run)

docker start container_name

You can now test from outside or from inside the container. If from inside:

/opt/webservice # wget localhost:5000/carles
Connecting to localhost:5000 (127.0.0.1:5000)
carles               100% |**************************************************************************************************************|     3  0:00:00 ETA
/opt/webservice # cat debug.log
2020-05-06 20:46:24,349 Initialized...
2020-05-06 20:46:24,359  * Running on http://0.0.0.0:5000/ (Press CTRL+C to quit)
2020-05-06 20:46:24,360  * Restarting with stat
2020-05-06 20:46:24,764 Initialized...
2020-05-06 20:46:24,771  * Debugger is active!
2020-05-06 20:46:24,772  * Debugger PIN: 123-456-789
2020-05-07 13:18:43,890 127.0.0.1 - - [07/May/2020 13:18:43] "GET /carles HTTP/1.1" 200 -

if you don’t use YAML files or what you need is to change the code, all this can be avoided as when you update the Python code, Flash realizes that and reloads. See this line in the logs:

2020-05-07 13:18:40,431  * Detected change in '/opt/webservice/wwebservice.py', reloading

The webservice.py autoreloads because we init Flask with debug set to on.

You can also start a container with shell directly:

sudo docker run -it ctop /bin/bash

The Ethernet standards group announces a new 800 GbE specification

Here is the link to the new: https://www.pcgamer.com/amp/the-ethernet-standards-group-developed-a-new-speed-so-fast-it-had-to-change-its-name/

This is a great new for scaling performance in the Data Centers. For routers, switches…

And this makes me think about all the Architects that are using Memcached and Redis in different Servers, in Networks of 1Gbps and makes me want to share with you what a nonsense, is often, that.

So the idea of having Memcache or Redis is just to cache the queries and unload the Database from those queries.

But 1Gbps is equivalent to 125MB (Megabytes) per second.

Local RAM Memory in Servers can perform at 24GB and more (24,000,000 Megabytes) per second, even more.

A PCIE NVMe drive at 3.5GB per second.

A local SSD drive without RAID 550 MB/s.

A SSD in the Cloud, varies a lot on the provider, number of drives, etc… but I’ve seen between 200 MB/s and 2.5GB/s aggregated in RAID.

In fact I have worked with Servers equipped with several IO Controllers, that were delivering 24GB/s of throughput writing or reading to HDD spinning drives.

If you’re in the Cloud. Instead of having 2 Load Balancers, 100 Front Web servers, with a cluster of 5 Redis with huge amount of RAM, and 1 MySQL Master and 1 Slave, all communicating at 1Gbps, probably you’ll get a better performance having the 2 LBs, and 11 Front Web with some more memory and having the Redis instance in the same machine and saving the money of that many small Front and from the 5 huge dedicated Redis.

The same applies if you’re using Docker or K8s.

Even if you just cache the queries to drive, speed will be better than sending everything through 1 Gbps.

This will matter for you if your site is really under heavy load. Most of the sites just query the MySQL Server using 1 Gbps lines, or 2 Gbps in bonding, and that’s enough.

CTOP.py

For updated information visit the main page for CTOP.py

Current stable version is v.0.8.9 updated on 2022-07-03.

Current branch under development is v.0.8.10 updated on 2022-07-03.

Version 0.8.0 added compatibility with Python 2, for older Systems.

Find the source code in: https://gitlab.com/carles.mateo/ctop

Clone it with:

git clone https://gitlab.com/carles.mateo/ctop.git

ctop.py is an Open Source tool for Linux System Administration that I’ve written in Python3. It uses only the System (/proc), and not third party libraries, in order to get all the information required.
I use only this modules, so it’s ideal to run in all the farm of Servers and Dockers:

  • os
  • sys
  • time
  • shutil (for getting the Terminal width and height)

The purpose of this tool is to help to troubleshot and to identify problems with a single view to a single tool that has all the typical indicators.

It provides in a single view information that is typically provided by many programs:

  • top, htop for the CPU usage, process list, memory usage
  • meminfo
  • cpuinfo
  • hostname
  • uptime
  • df to see the free space in / and the free inodes
  • iftop to see real-time bandwidth usage
  • ip addr list to see the main Ip for the interfaces
  • netstat or lsof to see the list of listening TCP Ports
  • uname -a to see the Kernel version

Other cool things it does is:

  • Identifying if you’re inside an Amazon VM, Google GCP, OpenStack VMs, Virtual Box VMs, Docker Containers or lxc.
  • Compatible with Raspberry Pi (tested on 3 and 4, on Raspbian and Ubuntu 20.04LTS)
  • Uses colors, and marks in yellow the warnings and in red the errors, problems like few disk space reaming or high CPU usage according to the available cores and CPUs.
  • Redraws the screen and adjust to the size of the Terminal, bigger terminal displays more information
  • It doesn’t use external libraries, and does not escape to shell. It reads everything from /proc /sys or /etc files.
  • Identifies the Linux distribution
  • Supports Plugins loaded on demand.
  • Shows the most repeated binaries, so you can identify DDoS attacks (like having 5,000 apache instances where you have normally 500 or many instances of Python)
  • Indicates if an interface has the cable connected or disconnected
  • Shows the Speed of the Network Connection (useful for Mellanox cards than can operate and 200Gbit/sec, 100, 50, 40, 25, 10…)
  • It displays the local time and the Linux Epoch Time, which is universal (very useful for logs and to detect when there was an issue, for example if your system restarted, your SSH Session would keep latest Epoch captured)
  • No root required
  • Displays recent errors like NFS Timed outs or Memory Read Errors.
  • You can enforce the output to be in a determined number of columns and rows, for data scrapping.
  • You can specify the number of loops (1 for scrapping, by default is infinite)
  • You can specify the time between screen refreshes, for long placed SSH sessions
  • You can specify to see the output in b/w or in color (default)

Plugins allow you to extend the functionality effortlessly, without having to learn all the code. I provide a Plugin sample for starting lights on a Raspberry Pi, depending on the CPU Load, and playing a message “The system is healthy” or “Warning. The CPU is at 80%”.

Limitations:

  • It only works for Linux, not for Mac or for Windows. Although the idea is to help with Server’s Linux Administration and Troubleshot, and Mac and Windows do not have /proc
  • The list of process of the System is read every 30 seconds, to avoid adding much overhead on the System, other info every second
  • It does not run in Python 2.x, requires Python 3 (tested on 3.5, 3.6, 3.7, 3.8, 3.9)

I decided to code name the version 0.7 as “Catalan Republic” to support the dreams and hopes and democratic requests of the Catalan people, to become and independent republic.

I created this tool as Open Source and if you want to help I need people to test under different versions of:

  • Atypical Linux distributions

If you are a Cloud Provider and want me to implement the detection of your VMs, so the tool knows that is a instance of the Amazon, Google, Azure, Cloudsigma, Digital Ocean… contact me through my LinkedIn.

Monitoring an Amazon Instance, take a look at the amount of traffic sent and received

Some of the features I’m working on are parsing the logs checking for errors, kernel panics, processed killed due to lack of memory, iscsi disconnects, nfs errors, checking the logs of mysql and Oracle databases to locate errors

Resources for Microservices and Business Domain Solutions for the Cloud Architect / Microservices Architect

First you have to understand that Python, Java and PHP are worlds completely different.

In Python you’ll probably use Flask, and listen to the port you want, inside Docker Container.

In PHP you’ll use a Frameworks like Laravel, or Symfony, or Catalonia Framework (my Framework) :) and a repo or many (as the idea is that the change in one microservice cannot break another it is recommended to have one git repo per Service) and split the requests with the API Gateway and Filters (so /billing/ goes to the right path in the right Server, is like rewriting URLs). You’ll rely in Software to split your microservices. Usually you’ll use Docker, but you have to add a Web Server and any other tools, as the source code is not packet with a Web Server and other Dependencies like it is in Java Spring Boot.

In Java you’ll use Spring Cloud and Spring Boot, and every Service will be auto-contained in its own JAR file, that includes Apache Tomcat and all other Dependencies and normally running inside a Docker. Tcp/Ip listening port will be set at start via command line, or through environment. You’ll have many git repositories, one per each Service.

Using many repos, one per Service, also allows to deploy only that repository and to have better security, with independent deployment tokens.

It is not unlikely that you’ll use one language for some of your Services and another for other, as well as a Database or another, as each Service is owner of their data.

In any case, you will be using CI/CD and your pipeline will be something like this:

  1. Pull the latest code for the Service from the git repository
  2. Compile the code (if needed)
  3. Run the Unit and Integration Tests
  4. Compile the service to an executable artifact (f.e. Java JAR with Tomcat server and other dependencies)
  5. Generate a Machine image with your JAR deployed (for Java. Look at Spotify Docker Plugin to Docker build from Maven), or with Apache, PHP, other dependencies, and the code. Normally will be a Docker image. This image will be immutable. You will probably use Dockerhub.
  6. Machine image will be started. Platform test are run.
  7. If platform tests pass, the service is promoted to the next environment (for example Dev -> Test -> PreProd -> Prod), the exact same machine is started in the next environment and platform tests are repeated.
  8. Before deploying to Production the new Service, I recommend running special Application Tests / Behavior-driven. By this I mean, to conduct tests that really test the functionality of everything, using a real browser and emulating the acts of a user (for example with BeHat, Cucumber or with JMeter).
    I recommend this specially because Microservices are end-points, independent of the implementation, but normally they are API that serve to a whole application. In an Application there are several components, often a change in the Front End can break the application. Imagine a change in Javascript Front End, that results in a call a bit different, for example, with an space before a name. Imagine that the Unit Tests for the Service do not test that, and that was not causing a problem in the old version of the Service and so it will crash when the new Service is deployed. Or another example, imagine that our Service for paying with Visa cards generates IDs for the Payment Gateway, and as a result of the new implementation the IDs generated are returned. With the mocked objects everything works, but when we deploy for real is when we are going to use the actual Bank Payment. This is also why is a good idea to have a PreProduction environment, with PreProduction versions of the actual Services we use (all banks or the GDS for flights/hotel reservation like Galileo or Amadeus have a Test, exactly like Production, Gateway)

If you work with Microsoft .NET, you’ll probably use Azure DevOps.

We IT Engineers, CTOs and Architects, serve the Business. We have to develop the most flexible approaches and enabling the business to release as fast as their need.

Take in count that Microservices is a tool, a pattern. We will use it to bring more flexibility and speed developing, resilience of the services, and speed and independence deploying. However this comes at a cost of complexity.

Microservices is more related to giving flexibility to the Business, and developing according to the Business Domains. Normally oriented to suite an API. If you have an API that is consumed by third party you will have things like independence of Services (if one is down the others will still function), gradual degradation, being able to scale the Services that have more load only, being able to deploy a new version of a Service which is independent of the rest of the Services, etc… the complexity in the technical solution comes from all this resilience, and flexibility.

If your Dev Team is up to 10 Developers or you are writing just a CRUD Web Application, a PoC, or you are an Startup with a critical Time to Market you probably you will not want to use Microservices approach. Is like killing flies with laser cannons. You can use typical Web services approach, do everything in one single Https request, have transactions, a single Database, etc…

But if your team is 100 Developer, like a big eCommerce, you’ll have multiple Teams between 5 and 10 Developers per Business Domain, and you need independence of each Service, having less interdependence. Each Service will own their own Data. That is normally around 5 to 7 tables. Each Service will serve a Business Domain. You’ll benefit from having different technologies for the different needs, however be careful to avoid having Teams with different knowledge that can have hardly rotation and difficult to continue projects when the only 2 or 3 Devs that know that technology leave. Typical benefit scenarios can be having MySql for the Billing Services, but having NoSQL Database for the image catalog, or to store logs of account activity. With Microservices, some services will be calling other Services, often asynchronously, using Queues or Streams, you’ll have Callbacks, Databases for reading, you’ll probably want to have gradual and gracefully failure of your applications, client load balancing, caches and read only databases/in-memory databases… This complexity is in order to protect one Service from the failure of others and to bring it the necessary speed under heavy load.

Here you can find a PDF Document of the typical resources I use for Microservice Projects.

You can also download it from my github repository:

https://github.com/carlesmateo/awesome-microservices

Do you use other solutions that are not listed?. Leave a message. I’ll investigate them and update the Document, to share with the Community.

Update 2020-03-06: I found this very nice article explaining the same. Microservices are not for everybody and not the default option: https://www.theregister.co.uk/AMP/2020/03/04/microservices_last_resort/

Update 2020-03-11: Qcom with 1,600 microservices says that microservices architecture is the las resort: https://www.theregister.co.uk/AMP/2020/03/09/monzo_microservices/