Author: Lisa

On the climate

I liked the format of CNN’s climate town hall event. The successful way of allowing candidates to convey a lot of information was basically an accident of the DNC’s stubbornness — they didn’t want to host a debate on climate change (why?!?!), and they own any situation where more than one candidate appears on a stage and talks (or something like that). So the only option available was to provide each candidate a chunk of time — *not* a debate. Sucks that it wasn’t freely available like the debates were, but we signed up for a free trial of YouTube TV to watch it. And we’ll get a free trial of something else for MSNBC’s climate debate later this month (Weld is supposed to be on the schedule).

I’m glad CNN had large time blocks for each candidate — they could have given everyone ten minutes and not had time for follow-up. And it was interesting to see Buttigieg use another approach to the question of why environmentalism is important — because ‘existential threat’ is not likely to convert any minds. But couching it in terms of our stewardship of God’s creations — that actually has potential to appeal to people who don’t care about species going extinct, habitat loss, clean air, clean water, their kids future. What does God think of how you are treating his creations? I cannot imagine that sort of digression coming into a debate format. There were a few forays into a nuclear discussion too — Yang’s liquid fluoride thorium reactor investment, Sanders assertion that we’ve got enough radioactive waste already. I’d have liked to see someone offering to invest in reactors fueled with used rods — not as part of our overall energy strategy, but because *something* needs to be done with the existing waste.

Possibly due to the origin of the event, possibly just demographics … but they had some good questions too. I loved seeing food policy repeatedly brought up as a component of climate change mitigation. And some realization that it’s not as easy as picking up a guy from a Gulf oil rig, sending him to classes for a few weeks, and dropping him off at his job on the solar farm. People get their identities tied up in their job — “what they do” — and that’s just as important as addressing the training and logistics of training people for new jobs.

It was great to see Booker acknowledging that he doesn’t know much about geoengineering and would have to research it before having an opinion. That’s a reasonable order of things. And I liked that Yang brought up geoengineering as a component of the solution — geoengineering is something I’d researched from the ML / data modeling side, and it was good to see it put forth as something other than nuc’ing hurricanes. I’d have liked to see the problems with capitalist enterprise being the source of technological solutions (although Yang touches on the issue in ditching GDP as the sole measurement of economic success).

Biden’s whole performance was expectedly underwhelming. I’m not sure where he got his rep as a great people-person politician … because WTF is the point of arguing about whose state is getting it worse? Biden’s digression into base readiness is something I wish had follow-up. He started down a path, then drew back saying something like “I cannot get into that”. Which made me wonder what exactly *that* was. He might have been talking about the DoD Climate Report from Jan 2019 and not wanted to get into it as a digression wasting his free airtime. But my mind went down the path … a VP’s got high level clearance. Does he keep it after leaving office? Can he still request briefings from DNI? Can he *not* get into it?

Cooper had a card about Andrew Goldman because they knew the question was going to be asked — it was basically a “gotcha” setup. But the exchange highlighted how Biden is the embodiment of what people hate about politicians. The “depends on what the meaning of the word ‘is’ is” legal technicality dodge. My takeaway is that Biden knew the guy had been a co-founder of a natural gas extraction company and runs a hedge fund with diversified investments. But the dude *isn’t* the CEO or active board member of a 100% fossil fuel company, so I’m following the letter of the agreement. How’d he make his first million? How’d he make his subsequent hundreds of millions? Not relevant. The alternative is that Biden’s campaign is inept, and they don’t do background research on people hosting high-dollar fundraising events. I believe that about as much as I believe a former model didn’t put any thought into the words on the back of the jacket she was wearing.

Other candidates did better jobs answering ‘gotcha’ questions. Sanders refused to pledge that no taxpayer will shoulder the 16 trillion dollar burden for his environmental plan (and I love the idea of TVA’ing renewable energy production, which I hadn’t heard before) because *some* taxpayers WILL pay more. Yang’s response about electric cars was great – it would have been a great place to talk about Porsche’s Taycan (another pure electric car with sub three-second 0-60 time) event earlier in the day. But telling people that they’re going to LOVE driving an electric car is spot on. Warren won’t take the light-bulb bait and wants to focus on the oil industry, the electric power industry and the building industry (although it would have been nice if *she* had enumerated them)

Some didn’t seem well prepared for obvious gotcha’s. Booker gets a bit cornered by the “taking away my burgers” question … because he’s a vegan and doesn’t want to seem like one of *those* vegans? Klobuchar doesn’t have a good response about the dairy or cattle industry because she may need to run again in Minnesota. But the Biden exchange seemed like an ambush in that it wasn’t an obvious question. Cooper returning to the Goldman topic for a closing clarification was bad for Biden too. The last impression you got of Biden was him being defensive about possibly taking money from an LNG guy. Which at least gives the appearance that Biden is losing his de facto choice status. Could you have seen CNN going after Clinton like that last time around?

The whole thing got me thinking that the next president has a chance of making some progress in spite of Moscow Mitch. Piece-meal some of it through budget reconciliation. There’s SCOTUS precedent that the Executive branch can take money from one place and allocate it elsewhere for something that has been intentionally not funded by Congress (yet another point where Trump probably doesn’t even understand the ramifications of his short-term ‘win’). Withhold subsidies from one place (fossil fuels, meat, dairy) and use that money toward other initiatives (increased solar/wind rebates, increased veggie subsidies, electric car rebates). Adjust dietary guidelines and national school lunch requirements (it’s be a huge uproar, but imagine the impact of schools doing meatless Monday). Sue the companies that have internal documents from *decades* ago indicating that they knew burning fossil fuels was environmentally destructive. It’s no different than the tobacco companies — well, it’s worse because even if you lived on a remote island for your entire life and never encountered anyone who bought anything from ExxonMobile … you’re impacted by their products — so why *don’t* we have trillion dollar settlements from them funding the public solar/wind power company?

Accessing MobyVM

I needed to map an addition port into an existing Docker container. Now I know the right thing to do is to create a new container and do it right this time but GitLab’s container has problems running on the Windows Docker Desktop. Permission-based problems that I’m not particularly included to attempt to sort out just to run a simple sandbox. Which means I’d need to drop my config file back in place & recreate my sandbox projects. And since I’m using CI/CD variables which don’t export … recreating the sandbox projects is a bit of a PITA.

On Linux, I can fix this by editing the config.v2.json and hostconfig.json files … but this is Windows running a funky Hyper-V Linux. And it turns out you can access the files on this MobyVM.

docker run -it --rm --privileged --pid=host justincormack/nsenter1

Now I’m able to cd into /var/lib/docker/containers, find the full ID for my GitLab container and cd into it, and edit the two config files. If it is running, you need to stop the container prior to editing the config files.

config.v2.json — add the port to “ExposedPorts”

chStdin”:false,”AttachStdout”:false,”AttachStderr”:false,”ExposedPorts”:{“22/tcp”:{},”443/tcp”:{},”80/tcp”:{},”4567/tcp”:{}},”Tty”:fal…

hostconfig.json — add the port to “PortBindings”

ult”,”PortBindings”:{“22/tcp”:[{“HostIp”:””,”HostPort”:”22″}],”443/tcp”:[{“HostIp”:””,”HostPort”:”443″}],”80/tcp”:[{“HostIp”:””,”HostPort”:”80″}],”4567/tcp”:[{“HostIp”:””,”HostPort”:”4567″}]},”Res…

 

Stop the Windows Docker service, start it, then start the container again. Voila! The new port for the container registry is there without recreating the container.

Using File System As A Git Repository

I’ve used GitLab for quite some time, and as a full featured CI/CD platform that also provides git functionality … it’s awesome. But it’s also serious overkill for someone who wants to coordinate code with another developer or two. Or just keep history for their code. Or a backup. To accomplish this, all you need is drive space. If you’ve got a file server, any folder on the share can be a git repository.

On the server, create a git repository and add something to it.

Z:\Temp\test>git init
Initialized empty Git repository in Z:/Temp/test/.git/

Z:\Temp\test>notepad testfile.txt

Z:\Temp\test>git add testfile.txt

Z:\Temp\test>git commit -m "Initial file upload"
[master (root-commit) 9a3ebe7] Initial file upload
1 file changed, 1 insertion(+)
create mode 100644 testfile.txt

Then on your client, either clone the repo from the drive path

C:\Users\lisa>git clone file://z:/temp/test
Cloning into 'test'...
remote: Enumerating objects: 3, done.
remote: Counting objects: 100% (3/3), done.
remote: Total 3 (delta 0), reused 0 (delta 0)
Receiving objects: 100% (3/3), done.

Or from the UNC path

C:\Users\lisa>git clone file://server01.example.com/data/temp/test
Cloning into 'test'...
remote: Enumerating objects: 3, done.
remote: Counting objects: 100% (3/3), done.
remote: Total 3 (delta 0), reused 0 (delta 0)
Receiving objects: 100% (3/3), done.

I prefer to use the UNC paths – if my drive letter fails to map for some reason, the UNC path is still available.

If you’ve got pre-existing code, there’s a bit of a different process. On the server, create an empty folder and use “git init” to initialize the empty repo. On the client where the code exists, run:

git init
git add *
git commit -m “Initial code upload”
git remote add origin git clone file://server01.example.com/data/temp/test
git push origin master

 

Finding Disabled Accounts In Active Directory

When using Active Directory (AD) as a source of user data, it’s useful to filter out disabled accounts. Unfortunately, AD has a lot of different security-related settings glomed together in the userAccountControl attribute. Which means there’s no single attribute/value combination you can use to ignore disabled accounts.

The decimal value you see for userAccountControl isn’t terribly useful, but display it in binary and each bit position has a meaning. The userAccountControl value is just the number with a bunch of bits set. Numbering the bits from left to right, here is what each one means.

Bit # Meaning
0 Unused – must be 0
1 Unused – must be 0
2 Unused – must be 0
3 Unused – must be 0
4 Unused – must be 0
5 ADS_UF_PARTIAL_SECRETS_ACCOUNT
6 ADS_UF_NO_AUTH_DATA_REQUIRED
7 ADS_UF_TRUSTED_TO_AUTHENTICATE_FOR_DELEGATION
8 ADS_UF_PASSWORD_EXPIRED
9 ADS_UF_DONT_REQUIRE_PREAUTH
10 ADS_UF_USE_DES_KEY_ONLY
11 ADS_UF_NOT_DELEGATED
12 ADS_UF_TRUSTED_FOR_DELEGATION
13 ADS_UF_SMARTCARD_REQUIRED
14 Unused – must be 0
15 ADS_UF_DONT_EXPIRE_PASSWD
16 Unused – must be 0
17 Unused – must be 0
18 ADS_UF_SERVER_TRUST_ACCOUNT
19 ADS_UF_WORKSTATION_TRUST_ACCOUNT
20 ADS_UF_INTERDOMAIN_TRUST_ACCOUNT
21 Unused – must be 0
22 ADS_UF_NORMAL_ACCOUNT
23 Unused – must be 0
24 ADS_UF_ENCRYPTED_TEXT_PASSWORD_ALLOWED
25 ADS_UF_PASSWD_CANT_CHANGE
26 ADS_UF_PASSWD_NOTREQD
27 ADS_UF_LOCKOUT
28 ADS_UF_HOMEDIR_REQUIRED
29 Unused – must be 0
30 ADS_UF_ACCOUNT_DISABLE
31 Unused – must be 0

Bit #30 indicates if the account is disabled — 1 if the account is disabled, 0 if the account is enabled. Simple and direct approach is to “and” the attribute value with 0b10 to extract just the bit we care about. When the and operation returns 0, the account is enabled. When it returns 2 (0x10), the account is disabled.

 

Inspecting An Element

In Firefox’s developer tools — instead of attempting to navigate through the HTML code to find the element, just right-click on it and select “Inspect Element”

You’ll get dropped into the Inspector tab right where you need to be.

There’s also an element picker tool that you can use instead — click on it & then click on the element within the page. Same result.

Upgrading Oracle SQL Developer

I’ve been using an Oracle database more in my new position … which means I’ve got the Oracle SQL Developer tool installed on my computer. My first upgrade was available yesterday … and it didn’t work. Not like threw an error, but like double click on the executable and nothing happens. It silently exits.

Turns out there’s something in appdata that needs to be cleared. I don’t run multiple versions of SQL Developer, so I could just blow away “%userprofile%\appdata\roaming\SQL Developer” and “%userprofile%\appdata\roaming\sqldeveloper” to clear whatever needs to be cleared. Click the icon and the program finally runs.

Setting up gitlab-runner

CLI on the GitLab server:

# Set up the GitLab Repo
curl -L https://packages.gitlab.com/install/repositories/runner/gitlab-runner/script.rpm.sh

# Install package
yum install gitlab-runner

# verify runner is executable
ll `which gitlab-runner`
# If needed, flag it executable – shouldn’t be a problem with RPM installations, but it’s been a problem for me with manual installs
#chmod ugo+x /usr/local/bin/gitlab-runner

# Register a runner
gitlab-runner register

# use URL & token from http://<GITLABSERVER>/admin/runners
# add tags, if you want to use tags to assign runner
# executor: shell (and docker, if you want to use docker executors. The shell executor is the simplest case, so I am starting there)

# start the runner
gitlab-runner start

On the GitLab Web GUI:

Admin section => Overview => Runners. Click pencil to edit the runner and uncheck “Lock to current projects”, and (unless you want to use tagging) check “Run untagged jobs”

** I was getting an error in every pipeline job saying the git command was not found.

For most other commands, you can append to the path in the before_script section of your .gitlab-ci.yml

before_script:
– export PATH=$PATH:/opt/whatever/odd/path/bin

But that doesn’t work in this case because we’re not getting that far: the bootstrap “stuff” cannot fetch the project to see the before script. Git, on my system, was part of the GitLab package. I just created a symlink into a “normal” binary location:

root@gitlab:~# which git
/opt/gitlab/embedded/bin/git
root@gitlab:~# ln -s /opt/gitlab/embedded/bin/git /usr/bin/git

And we’ve got successful test execution:

 

On Greenland

Trump wants to buy Greenland … which, yeah, it’s been suggested before. Way before, like 1946. And it wasn’t well received by the Danes at the time, so not exactly a stellar argument there. But Greenland isn’t a colony in the 1800’s style. A decade ago, the Act on Greenland Self-Government was granted. Which made Greenland’s parliament on par with the Danish parliament. Now, foreign policy and international agreements are still under Danish control, and there have been some power struggles in the intervening decade. But I don’t think anyone could buy Greenland from Denmark.

Software Flow Control and vim

In the early 90’s, one of the things I liked about Microsoft’s ecosystem was that they developed a standard for keyboard shortcuts. In most applications, developed by Microsoft or not, you could hit ctrl-p to print or ctrl-x to exit. Or ctrl-s to save. It’s quite convenient when I’m using Windows applications, but hitting ctrl-s to save without really thinking about it hangs vim. Hangs like “go into another shell and kill vim & that ssh session”. This is because ctrl-s, in Linux, means XOFF — the software flow control command that means “hi, I’m a thing from 1968 and my buffer is getting full. chill out for a bit and let me catch up”. Recovery is simple enough, send XON — “hi, that thing from 1968 again, and I’m all caught up. send me some more stuff”. That’s ctrl-q.

But there aren’t many slow anything‘s involved in computing these days, which means XON/XOFF isn’t the most useful of features for most people (* if you’ve got real serial devices attached … you may not be “most people” here). Instead of remembering ctrl-q gets gets vim back without killing it, just disable START/STOP control. Thing is it’s not really vim that’s using flow control — it’s the terminal emulator — so the “fix” isn’t something you’ll have to do in vim. In your ~/.bashrc or ~/.bash_profile (or globally in something like /etc/profile.d/disableFlowControl.sh)

# Disable XON/XOFF flow control so ctrl-s doesn’t hang vim
stty -ixon

If you can add -ixoff to avoid ctrl-q from meaning XON too … but I don’t bother since “start sending me data” doesn’t seem to hang anything.