Host foo.company.com
Port 22
Host *.company.com
Port 10022

Host *.company.com
Port 10022
Host foo.company.com
Port 22

• An AWS account with admin rights
• Full control over the domain name you use to send your newsletter
• A baremetal or virtual Linux Debian server you have the root access to
• NodeJS installed (8x and 10x are ok with Mailtrain)
• A MySQL/MariaDB instance with the root access to
• Redis 3x (not 5) if you want to use Redis (not mandatory)

• Configure AWS SES
• Configure your server by:
  • configuring your database
  • install Mailtrain
  • configuring your web server
• Configure your Mailtrain setup

Ask AWS SES to verify a domain

Generate the DKIM settings

AWS SES pending verification

AWS SES SMTP settings and credentials

MariaDB [(none)]> create database mailtrain;
Query OK, 1 row affected (0.00 sec)
MariaDB [(none)]> CREATE USER 'mailtrain' IDENTIFIED BY 'V3rYD1fF1cUlTP4sSW0rd!';
Query OK, 0 rows affected (0.01 sec)
MariaDB [(none)]> GRANT ALL PRIVILEGES ON mailtrain.* TO 'mailtrain'@localhost IDENTIFIED BY 'V3rYD1fF1cUlTP4sSW0rd!';
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> FLUSH PRIVILEGES;
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> show databases;
+--------------------+
  Database            
+--------------------+
  information_schema  
  mailtrain           
  mysql               
  performance_schema  
+--------------------+
6 rows in set (0.00 sec)
MariaDB [(none)]> Bye

[mysql]
host="localhost"
user="mailtrain"
password="V3rYD1ff1culT!"
database="mailtrain"

screen
NODE_ENV=production npm start

groupadd mailtrain
useradd -g mailtrain
chown -R mailtrain:mailtrain /var/www/mailtrain 

[Unit]
 Description=mailtrain
 After=network.target
[Service]
 Type=simple
 User=mailtrain
 WorkingDirectory=/var/www/mailtrain/
 Environment="NODE_ENV=production"
 Environment="PORT=3000"
 ExecStart=/usr/bin/npm run start
 TimeoutSec=15
 Restart=always
[Install]
 WantedBy=multi-user.target

systemctl daemon-reload
systemctl start mailtrain.service

map $http_upgrade $connection_upgrade  
  default upgrade;
  ''      close;
 
server  
  listen 80; 
  listen [::]:80;
  server_name mailtrain.toto.com;
  return 301 https://$host$request_uri;
 
server  
  listen 443 ssl;
  listen [::]:443 ssl;
  server_name mailtrain.toto.com;
  access_log /var/log/nginx/mailtrain.toto.com.access.log;
  error_log /var/log/nginx/mailtrain.toto.com.error.log;
  ssl_protocols TLSv1.2;
  ssl_ciphers EECDH+AESGCM:EECDH+AES;
  ssl_ecdh_curve prime256v1;
  ssl_prefer_server_ciphers on; 
  ssl_session_cache shared:SSL:10m;
  ssl_certificate     /etc/letsencrypt/live/mailtrain.toto.com/fullchain.pem;
  ssl_certificate_key /etc/letsencrypt/live/mailtrain.toto.com/privkey.pem;
  keepalive_timeout    70; 
  sendfile             on;
  client_max_body_size 0;
  root /var/www/mailtrain;
  location ~ /\.well-known\/acme-challenge  
    allow all;
   
  gzip on; 
  gzip_disable "msie6";
  gzip_vary on; 
  gzip_proxied any;
  gzip_comp_level 6;
  gzip_buffers 16 8k; 
  gzip_http_version 1.1;
  gzip_types text/plain text/css application/json application/javascript text/xml application/xml application/xml+rss text/javascript;
  add_header Strict-Transport-Security "max-age=31536000";
  location /   
    try_files $uri @proxy;
   
  location @proxy  
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Proto https;
    proxy_pass http://127.0.0.1:3000;
   
 

systemctl start nginx

Mailtrain mailer setup

Mailtrain to throttle sending emails to AWS SES

• Carl Chenet s blog
• Twitter account
• Mastodon account

• the whole inactivation/reactivation process can be modelled by a simple reversible reaction: $$ \mathrm A \rightleftharpoons \mathrm I $$ A is the active form, I the inactive form;
• the forward rate constant is \(k_i(E)\) (dependent on potential) and the backward rate constant is \(k_a(E)\), also dependent on potential;
• the experiment is done in a series of 5 steps at 3 different potentials: \(E_0\) then \(E_1\) then \(E_2\) then \(E_1\) then, finally, \(E_0\);
• the enzyme is assumed to be fully active at the beginning of the first step;
• a single experiment is used to obtain the values of \(k_i\) and \(k_a\) for the three potentials (although not reliably for the value at \(E_0\)
• the current given by the active species depends on potential (and it is negative because the enzyme catalyzes a reduction), and the inactive species gives no current;
• in addition to the reversible reaction above, there is an irreversible, potential-dependent loss.

QSoas> cd
QSoas> load 27.oxw
QSoas> load 27-blanc.oxw
QSoas> S 1 0

QSoas> strip-if x<30 x>300

QSoas> fit-linear-kinetic-system /species=2 /steps=0,1,2,1,0

• The I_1 and I_2 are the currents for the species 1 and 2, so something for 1 (active form) and 0 for 2 (inactive form). Moreover, the parameters I_2_#0 (and _#1, _#2) should be fixed and not free (since we don't need to adjust a current for the inactive form).
• The k_11 and k_22 correspond to species-specific irreversible loss. It is generally best to leave them fixed to 0.
• k_12 is the formation of 2 (I) from 1 (A), and k_21 is the formation of A from I. Their values will be determined for the three conditions. The default values should work here.
• The k_loss parameters are the rates of irreversible loss that apply indiscriminately on all species (unlike k_11 and k_22). They are adjusted and ther default values should work too.
• alpha_1_0 and alpha_2_0 are the initial concentrations of species 1 and 2, so they should be fixed to 1 and 0.
• Last, the xstart_a and (_b, _c, _d and _e) correspond to the starting times for the steps, here, 0, 60, 120, 210 and 270.

Three Seagrass always made decisions wholly and entire. All at once. choosing information as her aptitudes. Choosing the position of cultural liaison to the Lsel Ambassador. Choosing to trust her. choosing to come here, to take this assignment entirely, completely, and without pausing to look to see how deep the water was that she was leaping into.

The News

I joined Canonical, this February, to work on Ubuntu full-time! \o/
Those who know, they know that this is really very exciting for me because Canonical has been a dream company for me, for real (more about this below!). And hey, this is my first job, ever, so all the more reason to be psyched about, isn t it? ^_^ P.S. Keep reading and we ll meet my squad really sooon!

The Story Being an undergrad student (batch 2017-2021), I ve been slightly worried during my last two semesters, naturally, thinking about how s it all gonna pan out and what will I be doing, et al, because I ve been seeing all my friends and batchmates getting placed in companies or going for masters or at least having some sort of plans for their future and I, on the other hand, was hopelessly clueless. :D Well, to be fair, I did Google Summer of Code twice, in 2019 and 2020, became a Debian Developer in 2019, been a part of GCI and Outreachy, contributed to over dozens of open-source projects, et al, et al. So I wasn t all completely hopeless but for sure was completely clueless , heh. And for full disclosure, I was only slightly panicking because firstly, I did get placed in several companies and secondly, I didn t really need a job immediately since I was already getting paid to work on Debian stuff by Freexian, which was good enough. :)
(and honestly, Freexian has my whole heart! - more on that later sometime.) But that s not the point. I was still confused and worried and my mom & dad, more so than anyone. Ugh. We were all figuring out and she asked me places that I was interested to work in. And whilst I wasn t clear about things I wanted to do (and still am!) but I was (very) clear about this and so I told her about Canonical and also did tell her that it s a bit too ambitious for me to think about it now so I ll probably apply after some experience or something. and as they say, the world works in mysterious ways and well, it did for me! So back during the Ruby sprints (Feb 20), Kanashiro, the guy ( ), mentioned that his team was hiring and has a vacant position but I won t be eligible since I was still in my junior year. It was since then I ve been actively praying for Cronus, the god of time, to wave his magic wand and align it in such a way that the next opening should be somewhere near my graduation. And guess what? IT HAPPENED! 9 months later, in November 20, Kanashiro told me his team is hiring yet again and that I could apply this time! Without much (since there was some ) delay, I applied and started asking all sorts of questions to Kanashiro. No words are enough for him, he literally helped me throughout the process; from referring me to answering all sorts of doubts I had! And roughly after 2 months of interviewing, et al, my ambitious dream did come true and I finalyyyy signed my contract! \o/
(the interview process and what went on during those 10 weeks is a story for later ;))

The Server Team! \o This position, which I didn t mention earlier, was for the Server Team which is a team of 15 people, working to make Ubuntu server the best! And as I tweeted sometime back, the team is absolutely lovely, super kind, and consists of the best of teammates one could possibly ask for! Here s a quick sneak peek into our weekly team meeting. Thanks to Rafael for taking such a lovely picture. And yes, the cat Luna is a part of our squad! And oh, did I mention that we re completely remote and distributed?
FUN FACT: Our team covers all the TZs, that is, at any point of time (during weekdays), you ll find someone or the other from the team around! \o/ Anyway, our squad, managed by Rick is divided into two halves: Squeaky Wheels and Table Flip. Cool names, right?
Squeaky Wheels does the distro side of stuff and consists of Christian, Andreas, Rafael, Robie, Bryce, Sergio, Kanashiro, Athos, and now myself as well! And OTOH, Table Flip consists of Dan, Chad, Paride, Lucas, James, and Grant. Even though I interact w/ Squeaky Wheels more (basically daily), each of my teammates is absolutely lovely and equally awesome! Whilst I ll talk more about things here in the upcoming months, this is it for now! If there s anything, in particular, you d like to know more about, let me know! And lastly, here s us vibing our way through, making Ubuntu server better, cause that s how we roll!

---

Until next time.
:wq for today.

The Debian Janitor is an automated system that commits fixes for (minor) issues in Debian packages that can be fixed by software. It gradually started proposing merges in early December. The first set of changes sent out ran lintian-brush on sid packages maintained in Git. This post is part of a series about the progress of the Janitor. The FOSS world uses a wide variety of different build tools; given a git repository or tarball, it can be hard to figure out how to build and install a piece of software. Humans will generally know what build tool a project is using when they check out a project from git, or they can read the README. And even then, the answer may not always be straightforward to everybody. For automation, there is no obvious place to figure out how to build or install a project.

1
2
3
4
5
6
7
8
9

% git clone https://github.com/dulwich/dulwich
% cd dulwich
% ogni --schroot=unstable-amd64-sbuild dist
 
Writing dulwich-0.20.21/setup.cfg
creating dist
Creating tar archive
removing 'dulwich-0.20.21' (and everything under it)
Found new tarball dulwich-0.20.21.tar.gz in /var/run/schroot/mount/unstable-amd64-sbuild-974d32d7-6f10-4e77-8622-b6a091857e85/build/tmpucazj7j7/package/dist.

1
2
3
4
5
6
7
8
9

% wget https://download.samba.org/pub/ldb/ldb-2.3.0.tar.gz
% tar xvfz ldb-2.3.0.tar.gz
% cd ldb-2.3.0
% ogni install --prefix=/tmp/ldb
 
+ install /tmp/ldb/include/ldb.h (from include/ldb.h)
 
Waf: Leaving directory  /tmp/ldb-2.3.0/bin/default'
'install' finished successfully (11.395s)

1
2
3
4
5
6

% wget  https://cpan.metacpan.org/authors/id/T/TO/TORU/XML-LibXML-LazyBuilder-0.08.tar.gz _ <https://cpan.metacpan.org/authors/id/T/TO/TORU/XML-LibXML-LazyBuilder-0.08.tar.gz> _
% tar xvfz XML-LibXML-LazyBuilder-0.08.tar.gz
Cd XML-LibXML-LazyBuilder-0.08
% ogni test
 

The Population Issue and its many facets Another couple of weeks passed. A Lot of things happening, lots of anger and depression in folks due to handling in pandemic, but instead of blaming they are willing to blame everybody else including the population. Many of them want forced sterilization like what Sanjay Gandhi did during the Emergency (1975). I had to share So Long, My son . A very moving tale of two families of what happened to them during the one-child policy in China. I was so moved by it and couldn t believe that the Chinese censors allowed it to be produced, shot, edited, and then shared worldwide. It also won a couple of awards at the 69th Berlin Film Festival, silver bear for the best actor and the actress in that category. But more than the award, the theme, and the concept as well as the length of the movie which was astonishing. Over a 3 hr. something it paints a moving picture of love, loss, shame, relief, anger, and asking for forgiveness. All of which can be identified by any rational person with feelings worldwide.

Girl child What was also interesting though was what it couldn t or wasn t able to talk about and that is the Chinese leftover men. In fact, a similar situation exists here in India, only it has been suppressed. This has been more pronounced more in Asia than in other places. One big thing in this is human trafficking and mostly women trafficking. For the Chinese male, that was happening on a large scale from all neighboring countries including India. This has been shared in media and everybody knows about it and yet people are silent. But this is not limited to just the Chinese, even Indians have been doing it. Even yesteryear actress Rupa Ganguly was caught red-handed but then later let off after formal questioning as she is from the ruling party. So much for justice. What is and has been surprising at least for me is Rwanda which is in the top 10 of some of the best places in equal gender. It, along with other African countries have also been in news for putting quite a significant amount of percentage of GDP into public healthcare (between 20-10%), but that is a story for a bit later. People forget or want to forget that it was in Satara, a city in my own state where 220 girls changed their name from nakusha or unwanted to something else and that had become a piece of global news. One would think that after so many years, things would have changed, the only change that has happened is that now we have two ministries, The Ministry of Women and Child Development (MoWCD) and The Ministry of Health and Welfare (MoHFW). Sadly, in both cases, the ministries have been found wanting, Whether it was the high-profile Hathras case or even the routine cries of help which given by women on the twitter helpline. Sadly, neither of these ministries talks about POSH guidelines which came up after the 2012 gangrape case. For both these ministries, it should have been a pinned tweet. There is also the 1994 PCPNDT Act which although made in 1994, actually functioned in 2006, although what happens underground even today nobody knows . On the global stage, about a decade ago, Stephen J. Dubner and Steven Levitt argued in their book Freakonomics how legalized abortion both made the coming population explosion as well as expected crime rates to be reduced. There was a huge pushback on the same from the conservatives and has become a matter of debate, perhaps something that the Conservatives wanted. Interestingly, it hasn t made them go back but go forward as can be seen from the Freakonomics site.

Climate Change Another topic that came up for discussion was repeatedly climate change, but when I share Shell s own 1998 Confidential report titled Greenhouse effect all become strangely silent. The silence here is of two parts, there probably is a large swathe of Indians who haven t read the report and there may be a minority who have read it and know what already has been shared with U.S. Congress. The Conservative s argument has been for it is jobs and a weak we need to research more . There was a partial debunk of it on the TBD podcast by Matt Farell and his brother Sean Farell as to how quickly the energy companies are taking to the coming change.

Health Budget Before going to Covid stories. I first wanted to talk about Health Budgets. From the last 7 years the Center s allocation for health has been between 0.34 to 0.8% per year. That amount barely covers the salaries to the staff, let alone any money for equipment or anything else. And here by allocation I mean, what is actually spent, not the one that is shared by GOI as part of budget proposal. In fact, an article on Wire gives a good breakdown of the numbers. Even those who are on the path of free markets describe India s health business model as a flawed one. See the Bloomberg Quint story on that. Now let me come to Rwanda. Why did I chose Rwanda, I could have chosen South Africa where I went for Debconf 2016, I chose because Rwanda s story is that much more inspiring. In many ways much more inspiring than that South Africa in many ways. Here is a country which for decades had one war or the other, culminating into the Rwanda Civil War which ended in 1994. And coincidentally, they gained independence on a similar timeline as South Africa ending Apartheid in 1994. What does the country do, when it gains its independence, it first puts most of its resources in the healthcare sector. The first few years at 20% of GDP, later than at 10% of GDP till everybody has universal medical coverage. Coming back to the Bloomberg article I shared, the story does not go into the depth of beyond-expiry date medicines, spurious medicines and whatnot. Sadly, most media in India does not cover the deaths happening in rural areas and this I am talking about normal times. Today what is happening in rural areas is just pure madness. For last couple of days have been talking with people who are and have been covering rural areas. In many of those communities, there is vaccine hesitancy and why, because there have been whatsapp forwards sharing that if you go to a hospital you will die and your kidney or some other part of the body will be taken by the doctor. This does two things, it scares people into not going and getting vaccinated, at the same time they are prejudiced against science. This is politics of the lowest kind. And they do it so that they will be forced to go to temples or babas and what not and ask for solutions. And whether they work or not is immaterial, they get fixed and property and money is seized. Sadly, there are not many Indian movies of North which have tried to show it except for oh my god but even here it doesn t go the distance. A much more honest approach was done in Trance . I have never understood how the South Indian movies are able to do a more honest job of story-telling than what is done in Bollywood even though they do in 1/10th the budget that is needed in Bollywood. Although, have to say with OTT, some baggage has been shed but with the whole film certification rearing its ugly head through MEITY orders, it seems two steps backward instead of forward. The idea being simply to infantilize the citizens even more. That is a whole different ball-game which probably will require its own space.

Vaccine issues One good news though is that Vaccination has started. But it has been a long story full of greed by none other than GOI (Government of India) or the ruling party BJP. Where should I start with. I probably should start with this excellent article done by Priyanka Pulla. It is interesting and fascinating to know how vaccines are made, at least one way which she shared. She also shared about the Cutter Incident which happened in the late 50 s. The response was on expected lines, character assassination of her and the newspaper they published but could not critique any of the points made by her. Not a single point that she didn t think about x or y. Interestingly enough, in January 2021 Bharati Biotech was supposed to be share phase 3 trial data but hasn t been put up in public domain till May 2021. In fact, there have been a few threads raised by both well-meaning Indians as well as others globally especially on twitter to which GOI/ICMR (Indian Council of Medical Research) is silent. Another interesting point to note is that Russia did say in its press release that it is possible that their vaccine may not be standard (read inactivation on their vaccines and another way is possible but would take time, again Brazil has objected, but India hasn t till date.) What also has been interesting is the homegrown B.1.617 lineage or known as double mutant . This was first discovered from my own state, Maharashtra and then transported around the world. There is also B.1.618 which was found in West Bengal and is same or supposed to be similar to the one found in South Africa. This one is known as Triple mutant . About B.1.618 we don t know much other than knowing that it is much more easily transferable, much more infectious. Most countries have banned flights from India and I cannot fault them anyway. Hell, when even our diplomats do not care for procedures to be followed during the pandemic then how a common man is supposed to do. Of course, now for next month, Mr. Modi was supposed to go and now will not attend the G7 meeting. Whether, it is because he would have to face the press (the only Prime Minister and the only Indian Prime Minister who never has faced free press.) or because the Indian delegation has been disinvited, we would never know.

A good article which shares lots of lows with how things have been done in India has been an article by Arundhati Roy. And while the article in itself is excellent and shares a bit of the bitter truth but is still incomplete as so much has been happening. The problem is that the issue manifests in so many ways, it is difficult to hold on. As Arundhati shared, should we just look at figures and numbers and hold on, or should we look at individual ones, for e.g. the one shared in Outlook India. Or the one shared by Dr. Dipshika Ghosh who works in Covid ICU in some hospital

Dr. Dipika Ghosh sharing an incident in Covid Ward

Interestingly as well, while in the vaccine issue, Brazil Anvisa doesn t know what they are doing or the regulator just isn t knowledgeable etc. (statements by various people in GOI, when it comes to testing kits, the same is an approver.)

ICMR/DGCI approving internationally validated kits, Press release.

Twitter In the midst of all this, one thing that many people have forgotten and seem to have forgotten that Twitter and other tools are used by only the elite. The reason why the whole thing has become serious now than in the first phase is because the elite of India have also fallen sick and dying which was not the case so much in the first phase. The population on Twitter is estimated to be around 30-34 million and people who are everyday around 20 odd million or so, which is what 2% of the Indian population which is estimated to be around 1.34 billion. The other 98% don t even know that there is something like twitter on which you can ask help. Twitter itself is exclusionary in many ways, with both the emoticons, the language and all sorts of things. There is a small subset who does use Twitter in regional languages, but they are too small to write anything about. The main language is English which does become a hindrance to lot of people.

Censorship Censorship of Indians critical of Govt. mishandling has been non-stop. Even U.S. which usually doesn t interfere into India s internal politics was forced to make an exception. But of course, this has been on deaf ears. There is and was a good thread on Twitter by Gaurav Sabnis, a friend, fellow Puneite now settled in U.S. as a professor.

Gaurav on Trump-Biden on vaccination of their own citizens

Now just to surmise what has been happened in India and what has been happening in most of the countries around the world. Most of the countries have done centralization purchasing of the vaccine and then is distributed by the States, this is what we understand as co-operative federalism. While last year, GOI took a lot of money under the shady PM Cares fund for vaccine purchase, donations from well-meaning Indians as well as Industries and trade bodies. Then later, GOI said it would leave the states hanging and it is they who would have to buy vaccines from the manufacturers. This is again cheap politics. The idea behind it is simple, GOI knows that almost all the states are strapped for cash. This is not new news, this I have shared a couple of months back. The problem has been that for the last 6-8 months no GST meeting has taken place as shared by Punjab s Finance Minister Amarinder Singh. What will happen is that all the states will fight in-between themselves for the vaccine and most of them are now non-BJP Governments. The idea is let the states fight and somehow be on top. So, the pandemic, instead of being a public health issue has become something of on which politics has to played. The news on whatsapp by RW media is it s ok even if a million or two also die, as it is India is heavily populated. Although that argument vanishes for those who lose their dear and near ones. But that just isn t the issue, the issue goes much more deeper than that Oxygen:12%
Remedisivir:12%
Sanitiser:12%
Ventilator:12%
PPE:18%
Ambulances 28% Now all the products above are essential medical equipment and should be declared as essential medical equipment and should have price controls on which GST is levied. In times of pandemic, should the center be profiting on those. States want to let go and even want the center to let go so that some relief is there to the public, while at the same time make them as essential medical equipment with price controls. But GOI doesn t want to. Leaders of opposition parties wrote open letters but no effect. What is sad to me is how Ambulances are being taxed at 28%. Are they luxury items or sin goods ? This also reminds of the recent discovery shared by Mr. Pappu Yadav in Bihar. You can see the color of ambulances as shared by Mr. Yadav, and the same news being shared by India TV news showing other ambulances. Also, the weak argument being made of not having enough drivers. Ideally, you should have 2-3 people, both 9-1-1 and Chicago Fire show 2 people in ambulance but a few times they have also shown to be flipped over. European seems to have three people in ambulance, also they are also much more disciplined as drivers, at least an opinion shared by an American expat.

Pappu Yadav, President Jan Adhikar Party, Bihar May 11, 2021

What is also interesting to note is GOI plays this game of Health is State subject and health is Central subject depending on its convenience. Last year, when it invoked the Epidemic and DMA Act it was a Central subject, now when bodies are flowing down the Ganges and pyres being lit everywhere, it becomes a State subject. But when and where money is involved, it again becomes a Central subject. The States are also understanding it, but they are fighting on too many fronts.

Snippets from Karnataka High Court hearing today, 13th March 2021

One of the good things is most of the High Courts have woken up. Many of the people on the RW think that the Courts are doing Judicial activism . And while there may be an iota of truth in it, the bitter truth is that many judges or relatives or their helpers have diagnosed and some have even died due to Covid. In face of the inevitable, what can they do. They are hauling up local Governments to make sure they are accountable while at the same time making sure that they get access to medical facilities. And I as a citizen don t see any wrong in that even if they are doing it for selfish reasons. Because, even if justice is being done for selfish reasons, if it does improve medical delivery systems for the masses, it is cool. If it means that the poor and everybody else are able to get vaccinations, oxygen and whatever they need, it is cool. Of course, we are still seeing reports of patients spending in the region of INR 50k and more for each day spent in hospital. But as there are no price controls, judges cannot do anything unless they want to make an enemy of the medical lobby in the country. A good story on medicines and what happens in rural areas, see no further than Laakhon mein ek.

Allahabad High Court hauling Uttar Pradesh Govt. for lack of Oxygen is equal to genocide, May 11, 2021

The censorship is not just related to takedown requests on twitter but nowadays also any articles which are critical of the GOI s handling. I have been seeing many articles which have shared facts and have been critical of GOI being taken down. Previously, we used to see 404 errors happen 7-10 years down the line and that was reasonable. Now we see that happen, days weeks or months. India seems to be turning more into China and North Korea and become more anti-science day-by-day

Fake websites Before going into fake websites, let me start with a fake newspaper which was started by none other than the Gujarat CM Mr. Modi in 2005 .

Gujarat Satya Samachar 2005 launched by Mr. Modi.

And if this wasn t enough than on Feb 8, 2005, he had invoked Official Secrets Act

Mr. Modi invoking Official Secrets Act, Feb 8 2005 Gujarat Samachar

The headlines were In Modi s regime press freedom is in peril-Down with Modi s dictatorship. So this was a tried and tested technique. The above information was shared by Mr. Urvish Kothari, who incidentally also has his own youtube channel. Now cut to 2021, and we have a slew of fake websites being done by the same party. In fact, it seems they started this right from 2011. A good article on BBC itself tells the story. Hell, Disinfo.eu which basically combats disinformation in EU has a whole pdf chronicling how BJP has been doing it. Some of the sites it shared are

Times of New York
Manchester Times
Times of Los Angeles
Manhattan Post
Washington Herald
and many more. The idea being take any site name which sounds similar to a brand name recognized by Indians and make fool of them. Of course, those of who use whois and other such tools can easily know what is happening. Two more were added to the list yesterday, Daily Guardian and Australia Today. There are of course, many features which tell them apart from genuine websites. Most of these are on shared hosting rather than dedicated hosting, most of these are bought either from Godaddy and Bluehost. While Bluehost used to be a class act once upon a time, both the above will do anything as long as they get money. Don t care whether it s a fake website or true. Capitalism at its finest or worst depending upon how you look at it. But most of these details are lost on people who do not know web servers, at all and instead think see it is from an exotic site, a foreign site and it chooses to have same ideas as me. Those who are corrupt or see politics as a tool to win at any cost will not see it as evil. And as a gentleman Raghav shared with me, it is so easy to fool us. An example he shared which I had forgotten. Peter England which used to be an Irish brand was bought by Aditya Birla group way back in 2000. But even today, when you go for Peter England, the way the packaging is done, the way the prices are, more often than not, people believe they are buying the Irish brand. While sharing this, there is so much of Naom Chomsky which comes to my mind again and again

Caste Issues I had written about caste issues a few times on this blog. This again came to the fore as news came that a Hindu sect used forced labor from Dalit community to make a temple. This was also shared by the hill. In both, Mr. Joshi doesn t tell that if they were volunteers then why their passports have been taken forcibly, also I looked at both minimum wage prevailing in New Jersey as a state as well as wage given to those who are in the construction Industry. Even in minimum wage, they were giving $1 when the prevailing minimum wage for unskilled work is $12.00 and as Mr. Joshi shared that they are specialized artisans, then they should be paid between $23 $30 per hour. If this isn t exploitation, then I don t know what is. And this is not the first instance, the first instance was perhaps the case against Cisco which was done by John Doe. While I had been busy with other things, it seems Cisco had put up both a demurrer petition and a petition to strike which the Court stayed. This seemed to all over again a type of apartheid practice, only this time applied to caste. The good thing is that the court stayed the petition. Dr. Ambedkar s statement if Hindus migrate to other regions on earth, Indian caste would become a world problem given at Columbia University in 1916, seems to be proven right in today s time and sadly has aged well. But this is not just something which is there only in U.S. this is there in India even today, just couple of days back, a popular actress Munmun Dutta used a casteist slur and then later apologized giving the excuse that she didn t know Hindi. And this is patently false as she has been in the Bollywood industry for almost now 16-17 years. This again, was not an isolated incident. Seema Singh, a lecturer in IIT-Kharagpur abused students from SC, ST backgrounds and was later suspended. There is an SC/ST Atrocities Act but that has been diluted by this Govt. A bit on the background of Dr. Ambedkar can be found at a blog on Columbia website. As I have shared and asked before, how do we think, for what reason the Age of Englightenment or the Age of Reason happened. If I were a fat monk or a priest who was privileges, would I have let Age of Enlightenment happen. It broke religion or rather Church which was most powerful to not so powerful and that power was more distributed among all sort of thinkers, philosophers, tinkers, inventors and so on and so forth.

Situation going forward I believe things are going to be far more complex and deadly before they get better. I had to share another term called Comorbidities which fortunately or unfortunately has also become part of twitter lexicon. While I have shared what it means, it simply means when you have an existing ailment or condition and then Coronavirus attacks you. The Virus will weaken you. The Vaccine in the best case just stops the damage, but the damage already done can t be reversed. There are people who advise and people who are taking steroids but that again has its own side-effects. And this is now, when we are in summer. I am afraid for those who have recovered, what will happen to them during the Monsoons. We know that the Virus attacks most the lungs and their quality of life will be affected. Even the immune system may have issues. We also know about the inflammation. And the grant that has been given to University of Dundee also has signs of worry, both for people like me (obese) as well as those who have heart issues already. In other news, my city which has been under partial lockdown since a month, has been extended for another couple of weeks. There are rumors that the same may continue till the year-end even if it means economics goes out of the window.There is possibility that in the next few months something like 2 million odd Indians could die

The above is a conversation between Karan Thapar and an Oxford Mathematician Dr. Murad Banaji who has shared that the under-counting of cases in India is huge. Even BBC shared an article on the scope of under-counting. Of course, those on the RW call of the evidence including the deaths and obituaries in newspapers as a narrative . And when asked that when deaths used to be in the 20 s or 30 s which has jumped to 200-300 deaths and this is just the middle class and above. The poor don t have the money to get wood and that is the reason you are seeing the bodies in Ganges whether in Buxar Bihar or Gajipur, Uttar Pradesh. The sights and visuals makes for sorry reading

Pandit Ranjan Mishra son on his father s death due to unavailability of oxygen, Varanasi, Uttar Pradesh, 11th May 2021.

For those who don t know Pandit Ranjan Mishra was a renowned classical singer. More importantly, he was the first person to suggest Mr. Modi s name as a Prime Ministerial Candidate. If they couldn t fulfil his oxygen needs, then what can be expected for the normal public.

Conclusion Sadly, this time I have no humorous piece to share, I can however share a documentary which was shared on Feluda . I have shared about Feluda or Prodosh Chandra Mitter a few times on this blog. He has been the answer of James Bond from India. I have shared previously about The Golden Fortress . An amazing piece of art by Satyajit Ray. I watched that documentary two-three times. I thought, mistakenly that I am the only fool or fan of Feluda in Pune to find out that there are people who are even more than me. There were so many facets both about Feluda and master craftsman Satyajit Ray that I was unaware about. I was just simply amazed. I even shared few of the tidbits with mum as well, although now she has been truly hooked to Korean dramas. The only solace from all the surrounding madness. So, if you have nothing to do, you can look up his books, read them and then see the movies. And my first recommendation would be the Golden Fortress. The only thing I would say, do not have high hopes. The movie is beautiful. It starts slow and then picks up speed, just like a train. So, till later. Update The Mass surveillance part I could not do justice do hence removed it at the last moment. It actually needs its whole space, article. There is so much that the Govt. is doing under the guise of the pandemic that it is difficult to share it all in one article. As it is, the article is big

deb https://apt.kubernetes.io/ kubernetes-xenial main

gpg --recv-key 6A030B21BA07F4FB
gpg --list-sigs 6A030B21BA07F4FB
gpg --export 6A030B21BA07F4FB   apt-key add -

apt install libvirt-clients libvirt-daemon-system qemu-kvm dnsmasq kubectl

apt install libvirt-clients libvirt-daemon-system qemu-system-x86 dnsmasq kubectl

# get overview of virsh commands
virsh help
# list domains
virsh --connect qemu:///system list
# list block devices a domain uses
virsh --connect qemu:///system domblklist minikube
# show stats on block device usage
virsh --connect qemu:///system domblkstat minikube hda
# list virtual networks
virsh --connect qemu:///system net-list
# list dhcp leases on a virtual network
virsh --connect qemu:///system net-dhcp-leases minikube-net
# list network filters
virsh --connect qemu:///system nwfilter-list
# list real network interfaces
virsh --connect qemu:///system iface-list

• [1] https://kubernetes.io/docs/tutorials/kubernetes-basics/
• [2] https://wiki.debian.org/KVM
• [3] https://kubernetes.io/docs/tasks/tools/install-minikube/

• [1] http://tinyurl.com/y7rutdur
• [2] https://en.wikipedia.org/wiki/Ad_hominem
• [3] https://tinyurl.com/yg68dtco

ffmpeg ... --filter_complex "some arg with spaces" ...

filterarg='--filter_complex "some arg with spaces"'
...
ffmpeg $someargs $filterarg ...

#!/bin/bash
for i in "$@" ; do
  echo "=== $i"
done

#!/bin/bash
# Variant 1 - NOT WORKING
myargs='-filter_complex "arg with space"'
show_arguments $myargs

=== -filter_complex
=== "arg
=== with
=== space"

#!/bin/bash
# Variant 2 - WORKING
declare -a myargs
myargs=(-filter_complex "arg with space")
show_arguments "$ myargs[@] "

=== -filter_complex
=== arg with space

• Principles in Machine Learning
• Principles in Machine Learning /2
• Occam s Razor Principle of Parsimony
• Simple Models
• Computational Learning Theory
• Probably Approximately Correct Learning (PAC Learning) /1
• Probably Approximately Correct Learning (PAC Learning) /2
• PAC Learnable Examples
• VC Dimension (Vapnik-Chervonenkis-Dimension )
• VC Dimension Example
• Error Bounds and the VC Dimension
• No Free Lunch
• No Free Lunch Theorem
• No Free Lunch Theorem Explanation
• Bias-Variance Tradeoff
• Bias-Variance Tradeoff
• Bias vs. Variance
• Bias-Variance Decomposition
• Bias-Variance Illustration
• Different Kinds of Bias
• Data Often Has Bias
• AI Can Be Sexist and Racist
• Relationships

• Correlation does not Imply Causation
• Correlation does not Imply Causation /2
• Correlation does not Imply Causation /3
• Correlation with Statistics Classes
• Multiple Testing Problem
• Bonferroni s Principle Multiple Testing Problem
• Multiple Testing Problem

• Overfitting
• Underfitting and Overfitting
• Overfitting Decision Tree
• Overfitting Due to Noise
• Overfitting Due to Insufficient Examples

• Curse of Dimensionality
• Combinatorial Explosion
• Concentration of Distances
• Data is in the Margins
• Illustration: Shrinking Hyperspheres
• Illustration: Shrinking Hyperspheres /2
• Effect on Search in High Dimensionality
• Summary

• Intrinsic Dimensionality
• Estimating Intrinsic Dimensionality
• Angle-Based Intrinsic Dimensionality Intuition
• Angle-Based Intrinsic Dimensionality (ABID) /2
• Consequences & Solutions

• Distance Functions
• Distances, Metrics and Similarities
• Distances, Metrics and Similarities /2
• Distance Functions
• Distance Functions /2
• Similarity Functions
• Distances for Binary Data
• Jaccard Coefficient for Sets
• Example Distances for Categorical Data
• Mahalanobis Distance
• Scaling & Normalization
• To Scale, or not to Scale?
• To Scale, or not to Scale? /2

• Classification
• Prediction Problems
• Classification: A Multi-Stage Process
• Classification Problem
• Example
• Process of Constructing a Model
• Process of Applying the Model

• Evaluation and Selection of Classifiers
• Quick Recap: Classification
• Classifier Evaluation: Confusion Matrix
• Classifier Evaluation: Accuracy and Error-Rate
• Precision, Recall, and F-measure
• Classifier Evaluation: Multi-Class Confusion Matrix
• Training Accuracy vs. Accuarcy on New Data
• The Need for Validation
• Holdout Validation
• Cross-Validation
• Bootstrap Validation
• Considerations for Selecting a Model

• Bayesian Classification
• Bayes Classification: Motivation
• Bayes Theorem: Review
• Optimal Bayes Classifier
• Na ve Bayes Classifier
• Probability Models for a Single Attribute
• Multivariate Gaussian Bayes Classification
• Na ve Bayes Classifier: Example
• Na ve Bayes Classifier: Computational Aspects
• Na ve Bayes Classifier: Comments & Discussion

• Nearest-Neighbor Classification
• Nearest Neighbor Classifier Motivation
• Nearest Neighbor Classifier: Foundations
• Nearest Neighbor Classifier: Example
• Nearest Neighbor Classification: Example
• Nearest Neighbor Decision Rules

• Nearest-Neighbor as Density estimation
• Nearest Neighbor Classification and Density Estimation
• Predicting with Kernel Density Estimation with k=1,3,5,15
• Error Probability of Nearest Neighbors
• Nearest Neighbor Regression
• Nearest-Neighbor Classification: Comments & Discussion

• Decision Tree Learning
• Example (Variant of a Dataset in )
• Decision Tree Example
• Decision Trees as Rule-based Systems
• Basic Notions
• Constructing a Decision Tree /1
• Visual Interpretation of Decision Trees on R
• Constructing a Decision Tree /2
• Decision Tree Classification: Example

• Decision Tree Splitting
• Split for Categorical Attributes
• Split for Numeric Attributes
• Best Split Example
• Quality Measures for Splits
• Measure of Impurity: Gini Index
• Gini-Index: Example
• Information Gain
• Information Gain: Example
• Information Gain: Gain-Ratio
• Gain-Ratio: Beispiel
• Classification Error
• Gini, Entropy and Classification Error
• Comparing Split Selection Measures
• Splits for Numerical Attributes

• Ensembles and Meta-Learning
• Ensembles and Meta-Learning
• Error-Rate of Ensembles
• Random Forests
• Boosting
• Random Forest Classification: Example
• Gradient Boosting Classification: Example

• Support Vector Machine Motivation
• Support Vector Machines
• Support Vector Machines /2
• Finding the Best Separating Hyperplane
• Maximum Margin Hyperplane

• Affine Hyperplanes and Scalar Products
• Affine Hyperplanes
• Scalar Products
• Affine Hyperplanes /2

• Maximum Margin Hyperplane
• A Na ve Attempt
• Support Vectors Separable Data
• Computing the Maximum Margin Hyperplane (MMH)
• Computing the Maximum Margin Hyperplane (MMH) /2
• Boundary of the Maximum Margin Hyperplane (MMH)
• Deriving the Primal SVM Optimization Problem

• Training Support Vector Machines
• Optimization Problem
• Karush-Kuhn-Tucker KKT Conditions
• Switching to the Dual Problem
• Classification with the Dual SVM
• Optimizing the i
• Optimizing SVMs
• Sequential Minimal Optimization
• Further Improvements

• Non-linear SVM and the Kernel Trick
• Nonlinear SVM
• Nonlinear SVM /2
• Kernel Functions
• Soft Margin SVM Classifier
• Soft Margin SVM Classifier /2
• Soft Margin SVM Classifier /3
• Soft Margin SVM Classifier /4

• SVM Extensions and Conclusions
• Separation of more than 2 Classes
• Support Vector Regression
• Support Vector Regression Optimization Problem
• Support Vector Regression Dual
• Support Vector Data Description (SVDD)
• SVDD Dual Problem
• Support Vector Clustering
• SVMs: Comments & Discussion

• Neural Networks
• Biological Background (Simplified)
• Biological Background (Simplified) /2

• Threshold Logic Units
• Threshold Logic Units (TLUs)
• Threshold Logic Units Example
• Geometric Interpretation of TLUs
• Exclusive-Or (XOR) Problem
• Exclusive-Or (XOR) Problem /2
• Exclusive-Or (XOR) Problem /3
• Universality of TLUs
• Mark I Perceptron

• General Artificial Neural Networks
• Simplifying Threshold Logic Units
• Weight Matrices
• From TLUs to Multilayer Perceptrons
• Some Activation Functions
• Some Activation Functions /2
• Some Activation Functions /3
• Some Activation Functions /4

• Learning Neural Networks with Backpropagation
• Basic Gradient Descent
• Stochastic Gradient Descent
• Learning Single-Layer Perceptrons
• Backpropagation
• Training with Backpropagation

• Deep Neural Networks
• Universal Approximation Theorem
• Deep vs. Wide Neural Networks
• High vs. Low Dimensionality
• (Early) Problems of Deep Learning
• Autoencoders
• Layer-wise Pre-Training of Deep Neural Networks
• Dropout Regularization
• Batch Normalization
• Choosing Activation Functions

• Convolutional Neural Networks
• Depth vs. Convolutional Kernel Size
• Increasing the Training Data

• Recurrent Neural Networks
• Recurrent Neural Networks (RNNs) on Sequences
• Recurrent Neural Networks (RNN)
• Long-Short Term Memory (LSTM)
• Further Developments

• Conclusion
• Other Classifiers
• Problems of Classification

• Cluster Analysis Introduction
• What is Clustering?
• What is Clustering? /2
• Applications of Clustering
• Basic Steps for Clustering

• Hierarchical Agglomerative Clustering
• Distance of Clusters
• AGNES Agglomerative Nesting
• AGNES Agglomerative Nesting /2
• Extracting Clusters from a Dendrogram
• Benefits and Limitations of HAC

• Accelerating Hierarchical Clustering
• Complexity of Hierarchical Clustering
• Anderberg s Caching
• AGNES vs. Anderberg , NNChain , SLINK
• Example: Hierarchical Clustering with Anderberg

• K-means Clustering
• The Sum of Squares Objective
• The Standard Algorithm (Lloyd s Algorithm)
• Non-determinism & Non-optimality
• Initialization
• Initialization /2
• Complexity of k-Means Clustering

• Accelerating k-Means Clustering
• k-Means++: Weighted Random Initialization
• Making k-means Faster
• Bounding the Distances Elkan and Hamerly
• Hamerly s k-means
• Example: k-Means Clustering with Hamerly s Algorithm
• Speedup with Hamerly, Elkan, and Exponion

• Limitations of k-Means Clustering
• Benefits and Drawbacks of k-Means
• Choosing the Optimum k for k-Means
• Limitations of k-Means

• Extensions of k-Means Clustering
• k-Means and Distances
• k-Means Minimizes Sum of Squares, not Euclidean Distance!
• k-Means Variations for Other Distances
• Spherical k-Means for Text Clustering
• Pre-processing and Post-processing

• Partitioning Around Medoids (k-Medoids)
• k-medoids Clustering
• Partitioning Around Medoids
• Algorithm: Partitioning Around Medoids
• Algorithm: Partitioning Around Medoids /2
• Change in TD
• Finding the Best Swap Faster
• k-Medoids, k-Means style
• Example for the Inferiority of k-Means Style k-Medoids

• Gaussian Mixture Modeling Introduction
• Expectation-Maximization in Clustering
• Fitting Multiple Gaussian Distributions
• Gaussian Mixture Modeling as E-M-Optimization
• Algorithm: EM Clustering
• Numerical Issues in GMM

• Gaussian Mixture Modeling Demo

• BIRCH and BETULA
• BIRCH Clustering
• BIRCH Clustering Features
• BIRCH Distances
• BIRCH CF-Tree
• BETULA Cluster Features
• BETULA Distance Computations
• Accelerating k-Means with BIRCH and BETULA
• Accelerating GMM with BETULA

• Motivation Density-Based Clustering
• Density-based Clustering: Core Idea

• Density-Based Clustering Fundamentals
• Density-based Clustering: Foundations
• Density-based Clustering: Foundations /2
• Density-based Clustering: Foundations /3
• Density-reachability and Density-connectivity
• Density-reachability

• DBSCAN
• Clustering Approach
• Abstract DBSCAN Algorithm
• DBSCAN Algorithm
• DBSCAN Algorithm /2
• DBSCAN Algorithm /3
• DBSCAN in Context

• DBSCAN Parameterization
• Choosing DBSCAN parameters
• Choosing DBSCAN parameters /2
• Choosing DBSCAN parameters /3

• DBSCAN Extensions
• Generalized Density-based Clustering
• Grid-based Accelerated DBSCAN
• Anytime Density-Based Clustering (AnyDBC)
• Hierarchical DBSCAN* (HDBSCAN*)
• Improved DBSCAN Variations

• OPTICS Clustering
• Density-based Hierarchical Clustering
• Density-based Hierarchical Clustering /2
• OPTICS Clustering
• Cluster Order
• OPTICS Algorithm

• Cluster Extraction from OPTICS Plots
• OPTICS Reachability Plots
• Extracting Clusters from OPTICS Reachability Plots
• Role of the Parameters and minPts

• Understanding the OPTICS Cluster Order
• Properties of the OPTICS Cluster Order
• Cluster Order as Serialized Spanning Tree
• OPTICS as Density Spanning Trees
• Cluster Order to Dendrograms

• Spectral Clustering
• Minimum Cuts
• Graph Laplacian
• From Clustering Graphs to Clustering Data
• Spectral Clustering
• Spectral Clustering is Related to DBSCAN

• Biclustering and Subspace Clustering
• Biclustering & Subspace Clustering
• Bicluster Patterns
• Density-based Subspace Clustering
• Subspace Clustering with Apriori-Style Search
• Correlation Clustering
• 4C: Computing Correlation Connected Clusters
• Hough Transform
• CASH: Robust Clustering in Arbitrarily Oriented Subspaces

• Further Clustering Approaches
• CURE Clustering
• ROCK Clustering
• CHAMELEON
• Affinity Propagation Clustering
• Other Density-based Clustering Algorithms
• Further Clustering Approaches

May  2 16:38:37 server named[7372]: listening on IPv4 interface lo, 127.0.0.1#53
May  2 16:38:37 server named[7372]: listening on IPv4 interface eno4, 10.0.2.45#53
May  2 16:38:37 server named[7372]: listening on IPv4 interface eno4, 10.0.40.1#53
May  2 16:38:37 server named[7372]: listening on IPv4 interface eno4, 10.0.40.2#53
May  2 16:38:37 server named[7372]: listening on IPv4 interface eno4, 10.0.40.3#53
[...]
May  2 16:39:33 server named[7372]: listening on IPv4 interface eno4, 10.0.47.0#53
May  2 16:39:33 server named[7372]: listening on IPv4 interface eno4, 10.0.48.0#53
May  2 16:39:33 server named[7372]: listening on IPv4 interface eno4, 10.0.49.0#53
May  2 16:39:33 server named[7372]: listening on IPv6 interface lo, ::1#53
[...]
May  2 16:39:36 server named[7372]: zone localhost/IN: loaded serial 2
May  2 16:39:36 server named[7372]: all zones loaded
May  2 16:39:36 server named[7372]: running
May  2 16:39:36 server named[7372]: socket: file descriptor exceeds limit (123273/21000)
May  2 16:39:36 server named[7372]: managed-keys-zone: Unable to fetch DNSKEY set '.': not enough free resources
May  2 16:39:36 server named[7372]: socket: file descriptor exceeds limit (123273/21000)

OPTIONS="-u bind -S 150000"

listen-on   10.0.2.45;  ;

dig -t mx a704.example.com
socket.c:1740: internal_send: 10.0.2.45#53: Invalid argument
socket.c:1740: internal_send: 10.0.2.45#53: Invalid argument
socket.c:1740: internal_send: 10.0.2.45#53: Invalid argument
; <
> DiG 9.16.13-Debian <
> -t mx a704.example.com
;; global options: +cmd
;; connection timed out; no servers could be reached

Here is a sample of the strace output from tracing dig:
bind(20,  sa_family=AF_INET, sin_port=htons(0), 
sin_addr=inet_addr("0.0.0.0") , 16) = 0
recvmsg(20,  msg_namelen=128 , 0)       = -1 EAGAIN (Resource temporarily 
unavailable)
write(4, "\24\0\0\0\375\377\377\377", 8) = 8
sendmsg(20,  msg_name= sa_family=AF_INET, sin_port=htons(53), 
sin_addr=inet_addr("10.0.2.45") , msg_
namelen=16, msg_iov=[ iov_base="86\1 
\0\1\0\0\0\0\0\1\4a704\7example\3com\0\0\17\0\1\0\0)\20\0\0\0\0
\0\0\f\0\n\0\10's\367\265\16bx\354", iov_len=57 ], msg_iovlen=1, 
msg_controllen=0, msg_flags=0 , 0) 
= -1 EINVAL (Invalid argument)
write(2, "socket.c:1740: ", 15)         = 15
write(2, "internal_send: 10.0.2.45#53: Invalid argument", 45) = 45
write(2, "\n", 1)                       = 1
futex(0x7f5a80696084, FUTEX_WAIT_PRIVATE, 0, NULL) = 0
futex(0x7f5a80696010, FUTEX_WAKE_PRIVATE, 1) = 0
futex(0x7f5a8069809c, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f5a80698020, FUTEX_WAKE_PRIVATE, 1) = 1
sendmsg(20,  msg_name= sa_family=AF_INET, sin_port=htons(53), 
sin_addr=inet_addr("10.0.2.45") , msg_namelen=16, msg_iov=[ iov_base="86\1 
\0\1\0\0\0\0\0\1\4a704\7example\3com\0\0\17\0\1\0\0)\20\0\0\0\0\0\0\f\0\n\0\10's\367\265\16bx\354", 
iov_len=57 ], msg_iovlen=1, msg_controllen=0, msg_flags=0 , 0) = -1 EINVAL 
(Invalid argument)
write(2, "socket.c:1740: ", 15)         = 15
write(2, "internal_send: 10.0.2.45#53: Invalid argument", 45) = 45
write(2, "\n", 1)

Ubuntu bug #1702726 claims that an insufficient ARP cache was the cause of dig problems [3]. At the time I encountered the dig problems I was seeing lots of kernel error messages  neighbour: arp_cache: neighbor table overflow  which I solved by putting the following in /etc/sysctl.d/mine.conf:
net.ipv4.neigh.default.gc_thresh3 = 4096
net.ipv4.neigh.default.gc_thresh2 = 2048
net.ipv4.neigh.default.gc_thresh1 = 1024

Making that change (and having rebooted because I didn t need to run the server overnight) didn t entirely solve the problems. I have seen some DNS errors from Postfix since then but they are less common than before. When they happened I didn t have that error from dig. At this stage I m not certain that the ARP change fixed the dig problem although it seems likely (it s always difficult to be certain that you have solved a race condition instead of made it less common or just accidentally changed something else to conceal it). But it is clearly a good thing to have a large enough ARP cache so the above change is probably the right thing for most people (with the possibility of changing the numbers according to the required scale). Also people having that dig error should probably check their kernel message log, if the ARP cache isn t the cause then some other kernel networking issue might be related.
Preliminary Results
With Postfix I m seeing around 24,000 messages relayed per minute with more than 60% CPU time idle. I m not sure exactly how to count idle time when there are 12 CPU cores and 24 hyper-threads as having only 1 process scheduled for each pair of hyperthreads on a core is very different to having half the CPU cores unused. I ran my script to disable hyper-threads by telling the Linux kernel to disable each processor core that has the same core ID as another, it was buggy and disabled the second CPU altogether (better than finding this out on a production server). Going from 24 hyper-threads of 2 CPUs to 6 non-HT cores of a single CPU didn t change the thoughput and the idle time went to about 30%, so I have possibly halved the CPU capacity for these tasks by disabling all hyper-threads and one entire CPU which is surprising given that I theoretically reduced the CPU power by 75%. I think my focus now has to be on hyper-threading optimisation.
Since 2006 the performance has gone from ~20 messages per minute on relatively commodity hardware to 24,000 messages per minute on server equipment that is uncommon for home use but which is also within range of home desktop PCs. I think that a typical desktop PC with a similar speed CPU, 32G of RAM and SSD storage would give the same performance. Moore s Law (that transistor count doubles approximately every 2 years) is often misquoted as having performance double every 2 years. In this case more than 1024* the performance over 15 years means the performance doubling every 18 months. Probably most of that is due to SATA SSDs massively outperforming IDE hard drives but it s still impressive.
Notes
I ve been using example.com for test purposes for a long time, but RFC2606 specifies .test, .example, and .invalid as reserved top level domains for such things. On the next iteration I ll change my scripts to use .test.
My current test setup has a KVM virtual machine running my bhm program to receive mail which is taking between 20% and 50% of a CPU core in my tests so far. While that is happening the kvm process is reported as taking between 60% and 200% of a CPU core, so kvm takes as much as 4* the CPU of the guest due to the virtual networking overhead   even though I m using the virtio-net-pci driver (the most efficient form of KVM networking for emulating a regular ethernet card). I ve also seen this in production with a virtual machine running a ToR relay node.
I ve fixed a bug where Postal would try to send the SMTP quit command after encountering a TCP error which would cause an infinite loop and SEGV.

[1] https://doc.coker.com.au/papers/benchmarking-mail-relays-and-forwarders/
[2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=987927
[3] https://bugs.launchpad.net/ubuntu/+source/dnsmasq/+bug/1702726


Related posts:

interesting things  /tmp /mnt/bind bind bind 0 0 Today I discovered that...
new release of postal  Today I have released a significant new version of my...
BIND Stats  In Debian the BIND server will by default append statistics...

There was a boy named Eustace Clarence Scrubb and he almost deserved it.

"Oh Aslan," said she, "it was kind of you to come." "I have been here all the time," said he, "but you have just made me visible." "Aslan!" said Lucy almost a little reproachfully. "Don't make fun of me. As if anything I could do would make you visible!" "It did," said Aslan. "Did you think I wouldn't obey my own rules?"

"My son," said Ramandu, "it is not for you, a son of Adam, to know what faults a star can commit."

"My own plans are made. While I can, I sail east in the Dawn Treader. When she fails me, I paddle east in my coracle. When she sinks, I shall swim east with my four paws. And when I can swim no longer, if I have not reached Aslan's country, or shot over the edge of the world in some vast cataract, I shall sink with my nose to the sunrise and Peepiceek shall be the head of the talking mice in Narnia."

screenshot showing power consumption of server

• Friendica (MariaDB): 31 GB for 700 users
• Mastodon (PostgreSQL): 15 GB for 1000 users
• Matrix-Synapse (PostgreSQL): 5 GB for 1 user
• XMPP (PostgreSQL): 0.4 GB for 200 users

• Webserver VM (Matrix, Friendica & Mastodon): 13410130 s / 130%
• XMPP VM: 944275 s / 5.4%

• Friendica: 11575 lines
• Matrix: 8174 lines
• Mastodon: 3212 lines

• Reviewed and merged two pull request from Michael K. to my django-slack library which provides a convenient wrapper between projects using the Django web development framework and the Slack chat platform. Michael's pull requests made Python 3.6+ a hard requirement (#102) and to run the tests against Python 3.9 (#103).
• Made a "no-change" release of my django-email-from-template library to upload a Python wheel (.whl) file to PyPI. [...]

• As part of my role of being the assistant Secretary of the Open Source Initiative and a board director of Software in the Public Interest, I attended their respective monthly meetings. My term on the OSI board has been slightly extended due to the discovery of a vulnerability in OSI's recent election as a result, the 2021 election will be re-run to ensure transparency of the process.
• Filed a pull request against libsass-python (a Python implementation of the Sass CSS preprocessor) in order to make the build reproducible. (#319)

• Interviewed the Ford Foundation's Michael Brennan and published the transcript on our website.
• Drafted, published and publicised our monthly report.
• Filed an upstream pull request for libsass-python (a Python implementation of the Sass CSS preprocessor) in order to make the build reproducible. [...
• In Debian:
  • Kept isdebianreproducibleyet.com up to date. [...]
  • I also submitted 2 patches to fix specific reproducibility issues in pristine-lfs and rust-configparser.
• Categorised a large number of packages and issues in the Reproducible Builds notes.git" repository.

• Updated the main Reproducible Builds website and documentation:
  • Highlight our mailing list on the Contribute. page [...]
  • Add a noun (and drop an unnecessary full-stop) on the landing page. [...][...]
  • Correct a reference to the date metadata attribute on reports, restoring the display of months on the homepage. [...]
  • Correct a typo of "instalment" within a previous news entry. [...]
  • Added a conspicuous "draft" banner to unpublished blog posts in order to match the report draft banner. [...]

• I also made the following changes to diffoscope, including uploading versions 172 and 173 to Debian:
  • Add support for showing annotations in PDF files. (#249)
  • Move to the assert_diff helper in test_pdf.py. [...]

• redis (5:6.2.2-1) (to experimental) New upstream release.
• python-django:
  • 2.2.20-1 New upstream security release.
  • 3.2-1 (to experimental) New major upstream release (release notes).
• hiredis (1.0.0-2) Build with SSL/TLS support (#987114), and overhaul various aspects of the packaging.
• mtools (4.0.27-1) New upstream release.

• Investigated and triaged avahi (CVE-2021-3468), exiv2 (CVE-2021-3482), file-roller (CVE-2020-36314), fluidsynth (CVE-2021-28421), gnuchess (CVE-2021-30184), gpac (CVE-2021-28300), imagemagick (CVE-2021-20309, CVE-2021-20243), ircii (CVE-2021-29376), jetty9 (CVE-2021-28163), libcaca (CVE-2021-30498, CVE-2021-30499), libjs-handlebars, libpano13, libpodofo (CVE-2021-30469, CVE-2021-30470, CVE-2021-30471, CVE-2021-30472), mediawiki, mpv (CVE-2021-30145), nettle (CVE-2021-20305), nginx (CVE-2020-36309), nim (CVE-2021-21372, CVE-2021-21373, CVE-2021-21374), node-glob-parent (CVE-2020-28469), openexr (CVE-2021-3474), python-django-registration (CVE-2021-21416), qt4-x11 (CVE-2021-3481), qtsvg-opensource-src (CVE-2021-3481), ruby-kramdown, scrollz (CVE-2021-29376), syncthing (CVE-2021-21404), thunderbird (CVE-2021-23991, CVE-2021-23992, CVE-2021-23993) & wordpress (CVE-2021-29447).
• Issued DLA 2620-1 to address a cross-site scripting (XSS) vulnerability in python-bleach, a whitelist-based HTML sanitisation library.
• Issued DLA 2622-1 and ELA 402-1 as it was discovered that there was a potential directory traversal issue in Django, the popular Python-based web development framework. The vulnerability could have been exploited by maliciously crafted filenames. However, the upload handlers built into Django itself were not affected. (#986447)
• Jan-Niklas Sohn discovered that there was an input validation failure in the X.Org display server. Insufficient checks on the lengths of the XInput extension's ChangeFeedbackControl request could have lead to out of bounds memory accesses in the X server. These issues could have led to privilege escalation for authorised clients, particularly on systems where the X server is running as a privileged user. I, therefore, issued both DLA 2627-1 and ELA 405-1 to address this problem.
• Frontdesk duties, reviewing others' packages, participating in mailing list discussions, etc., as well as attending our monthly meeting.

• Abhijith PA did 9.0h (out of 9h assigned and 12h from February), thus carrying over 12h to April.
• Anton Gladky did 12h (out of 12h assigned).
• Ben Hutchings did 25.75h (out of 16h assigned and 12.25h from February), thus carrying over 2.5h to April.
• Chris Lamb did 18h (out of 18h assigned).
• Holger Levsen did 6h coordinating/managing the LTS team.
• Markus Koschany did 30h (out of 30h assigned).
• Ola Lundqvist did 6.h (out of 10h from February).
• Roberto C. S nchez did 9h (out of 32h assigned and 21.5h from February) and gave 40h back, thus carrying over 4.5h to April.
• Sylvain Beucler did 30h (out of 30h assigned).
• Thorsten Alteholz did 30h (out of 30h assigned).
• Utkarsh Gupta did 60h (out of 60h assigned).

• Platinum sponsors:
  • TOSHIBA (for 67 months)
  • GitHub (for 58 months)
  • Civil Infrastructure Platform (CIP) (for 35 months)
• Gold sponsors:
  • Blablacar (for 82 months)
  • Roche Diagnostics International AG (for 78 months)
  • Linode (for 72 months)
  • Babiel GmbH (for 61 months)
  • Plat Home (for 61 months)
  • University of Oxford (for 17 months)
  • Deveryware (for 4 months)
• Silver sponsors:
  • The Positive Internet Company (for 83 months)
  • Domeneshop AS (for 82 months)
  • Nantes M tropole (for 77 months)
  • Univention GmbH (for 68 months)
  • Universit Jean Monnet de St Etienne (for 68 months)
  • Ribbon Communications, Inc. (for 62 months)
  • Exonet B.V. (for 52 months)
  • Leibniz Rechenzentrum (for 46 months)
  • CINECA (for 35 months)
  • Minist re de l Europe et des Affaires trang res (for 30 months)
  • Cloudways Ltd (for 19 months)
  • Dinahosting SL (for 17 months)
  • Bauer Xcel Media Deutschland KG (for 11 months)
  • Platform.sh (for 11 months)
  • Moxa Intelligence Co., Ltd. (for 5 months)
  • sipgate GmbH
  • OVH US LLC
  • Tilburg University
• Bronze sponsors:
  • Evolix (for 83 months)
  • Seznam.cz, a.s. (for 83 months)
  • Intevation GmbH (for 80 months)
  • Linuxhotel GmbH (for 80 months)
  • Daevel SARL (for 78 months)
  • Bitfolk LTD (for 77 months)
  • Megaspace Internet Services GmbH (for 77 months)
  • NUMLOG (for 77 months)
  • Greenbone Networks GmbH (for 76 months)
  • WinGo AG (for 76 months)
  • Ecole Centrale de Nantes LHEEA (for 72 months)
  • Entr ouvert (for 67 months)
  • Adfinis AG (for 65 months)
  • GNI MEDIA (for 59 months)
  • Laboratoire LEGI UMR 5519 / CNRS (for 59 months)
  • Tesorion (for 59 months)
  • Bearstech (for 50 months)
  • LiHAS (for 50 months)
  • People Doc (for 47 months)
  • Catalyst IT Ltd (for 45 months)
  • Supagro (for 40 months)
  • Demarcq SAS (for 39 months)
  • Universit Grenoble Alpes (for 25 months)
  • TouchWeb SAS (for 17 months)
  • SPiN AG (for 14 months)
  • CoreFiling (for 9 months)
  • Institut des sciences cognitives Marc Jeannerod (for 4 months)
  • Observatoire des Sciences de l Univers de Grenoble

• What issues and areas to consider
• How other projects and foundations have approached these topics
• What FOSS foundations bring to the table
• How to choose a FOSS foundation

• Role and activities of foundations
• Challenges faced and gaps in the service offerings
• Operational aspects, including reasons for starting an org and choice of jurisdiction
• Trends, such as the "foundation in a foundation" model
• Recommendations for different stakeholders