DEV Community: Tim Roberts

Semester in Review

Tim Roberts — Fri, 22 Apr 2022 00:40:36 +0000

This is it, my final blog for the semester. It's been quite the ride. To start off I will briefly go over what I have done to contribute this past week for the release of 3.0. I will then do a reflection on the course and give some insights into my experience.

Release 3.0

This week was a pretty busy week, what with exams and final projects being due. Even still, I was able to get some things done before we ship 3.0.

I was able to land my Octokit PR that I have been plugging away at the last little bit. I ended up not only adding the Octokit package to the dependency-discovery, but I also did the same in our Status microservice since the throttling, and retry plugins are included in the main package. I also ended up adding throttling and retry functionality to the dependency-discovery since that was the entire point of updating the package.

That was my only PR for this week. I did however do some reviews to try and help others get stuff landed for the final release of the semester. I first did another review of Tuee's PR for removing the legacy backend. This PR is an important one. We want to get it landed for 3.0 so we can start using all the cool new Supabase stuff that has been added this semester. I had some issues testing it, but after clarifying with Tuee I was able to spin it up and approve it for merging.

The other two reviews I did this week where a little simpler to complete. The first one was a bug fix by Francesco. He just disabled some of the logging that was cluttering up the Docusaurus build log.

The other review was a quick patch fix made by the professor himself to try and get our CI tests to not consistently fail. This would make it so people can get stuff landed for the final release, but we would have to properly fix it at a later date.

Reflecting

Looking back on the course and I am pretty proud of my accomplishments. In the beginning I had set out to improve my skills with Docker and the backend stuff in general. I took on, within reason for the amount of work I was able to do, any and all issues related to Docker.

This eventually lead to me taking responsibility for, and 'owning' the launch of our Docker Registry. Even though I 'owned' it, a lot of work and help was provided by a number of others. Josue in particular was a great help since he has such great understanding of the project as a whole. I am especially appreciative of his help because he is not even a student this semester, all the help he gave was during his own free time.

A lot of what I did for getting the registry up and running was just getting things like PR's started to the best of my ability, and then post it as a draft PR and ask for help putting the last few pieces together. It was nice working in such a collaborative manor.

I can especially appreciate the collaboration because at my place of employment, I am the lone developer after a senior dev took another opportunity a few months ago. It's hard working through things as a junior developer by yourself. Having many minds on a single issue can not only get the task done sooner, but also save an individual from getting stressed and burnt out working on issues that are giving them troubles. This has shown me that I do not want to work at a start up any more, I would rather have a team I can depend on and break through issues with.

So, while a lot of the testing of the registry setup was done by Josue and Dave. I still made sure it all happened. During the triage and planning meetings, I would make sure to ask what direction, or what the next steps would be for the registry. I was driving the issues and making sure things got done by following up or starting thins on my own. That is why, even though I feel like I was guided through a lot of what had to get done code wise, I still feel like I can honestly say that I 'owned' the issue, and that I was responsible for getting the registry up and running.

One problem with me taking ownership of the registry is that I put all my attention to it and was not keeping up with what everyone else was doing. Yea, I knew the big picture things like adding Supabase, or getting the Parser service up and running. When the registry launched there was still about a month left of the course and when I was looking at issues to pick up and help with, it felt like I was so far behind. I felt that since other people had been working on things related to these issues that it would be far easier and take much less time than if I were to try and bash away at it.

A prime example of the is how long it took me to do the Octokit stuff. Jerry had mentioned it would be a relatively quick and simple PR to do when I took it over from him. Looking back on it, it was, but I had to do a bunch of research to just find out what Octokit was, and then I was uncertain of how and where we wanted to use the expanded package. Overall, it took me far longer to get the PR landed than it would have taken Jerry, but he had other, more important issues to try and tackle.

I tired to make up for this by getting in on more reviews, but most every time I went to check if something needed to be reviewed and everything was marked with changes requested so I left it alone. I also tried looking into some of the Supabase issues, but I haven't been following it at all so it was quite daunting to try and step in to fix things when I don't even know how to spin it up.

Wrapping up

Just like with the previous course, this course was another great example of how to give students real life work experience that will look good on most any resume. I feel much more confident in my Git abilities and feel like I have learned a lot about Docker, nginx, registries, and workflows/yaml files.

These two course should be core curriculum, but I understand why they are not. There is only one David Humphrey, and without him I find it hard to see these courses being as successful as they are. He really goes above and beyond what most of the other teachers do. I just wish I had known his cloud computing course was going to be available to my program this summer. If I had a time machine I would gladly take that course over the Azure Fundamentals cloud course I just completed this semester.

Close to the end

Tim Roberts — Sat, 16 Apr 2022 13:17:39 +0000

This weeks blog marks the second last blog I will be doing for this course. It's been a pretty wild ride, but I will do a proper reflection on the course for my last blog coming up next week. For now, as usual, let's talk about the week that was.

Octokit

In last weeks blog I had mentioned that I did some R&D into adding Octokit to the dependency discovery service, but needed to follow up with Jerry to clarify some things. After following up with Jerry, I figured out we just wanted to update the usage of octokit/request to be from octokit instead. This would give us more options later in the future.

After a couple of iterations, I was able to put up a PR for the issue. All I had to do was get ride of any references to octokit/request in the pnpm-lock.yaml and then do a pnpm install octokit. From there I just imported Octokit, assigned it to a variable and used the code as is, but just added the variable in front of the existing request like so.

During the triage meeting, we had some more discussion arise about the PR. We decided that it would be best to also update the usage of some other octokit stuff we have going on in our status microservice.

I ran into some issues trying to implement this so I switched my PR to a draft and am currently still troubleshooting. Should be able to get this in for 3.0.

Migrating with Prisma

I unintentionally ended up being the tester for the new Prisma migrations Duc had set up. During the planning meeting we had decided to meet at 2pm to properly triage some of the remaining issues for integrating the parser service. 2pm roles around and I notice that there is still a meeting going on that started at 1pm about the prisma migrations. When I joined the meeting they where just finishing up writing the documentation for performing a migration.

The professor then called me out to test it, which made sense because he wanted to see if the instructions they just made where clear enough for someone to successfully do a migration by just reading the documentation. Considering I had never even visited the Portainer site before, I was a great candidate. There where only a couple of things missing from the documentation, and a few of the steps needed to be rearranged to be in the proper order of operations, but I was able to successfully migrate! (Should have saved a screen grab of it)

Closing

So that was my week 13/14 for this course. As mentioned earlier, my last blog coming up next week will be a recap of the week that was, and then a overall reflection of how the course was and what I took away from it.

Second to None

Tim Roberts — Fri, 08 Apr 2022 22:14:56 +0000

Here we are, the release week of 3.0-Alpha. I did not get any PR's in this week, but I did help review some and did some research for others.

Gitpod bug

I made some progress on the draft PR I submitted last week for fixing Gitpod in the repository. I had incorporated the review suggestion, then pushed it to the PR and tried launching Gitpod again. It went through almost the entire process without a hiccup, but still exited with code 1 :(

It was the same issue that was rearing it's head before where it cannot connect to our Elasticserach instance. Originally the professor suggested adding a restart: unless-stopped line to the parser service that is trying to establish a connection. After adding that though the professor has a new idea as to why it's not working.

After looking into the issue I noticed that there was a couple of PR's related to it, so I decided to review them.

Review 1

This PR by Tuee was a doozy, 105 files changed!

Most of the files changed where straight up deletions, so there was not too much stuff to review. I looked through all the files and all appeared to be good on the surface level. I also spun it up locally and all seemed to be working. I left an approval of a review but noted that I am not fully aware of the greater implications of removing all of this code. I was the first review on the PR, but since then there have been a few others chime in and point out some issues that needed to be addressed. Good thing this is a team effort.

Review 2

I also did another review of A PR that Amasia had submitted. The PR added the ability to search by Title in Telescope's advanced search. It was a relatively straightforward PR to review. Spin the thing up and make sure the search results are relevant.

R&D

During our triage meeting on Tuesday, I had stepped up to take an issue off of Jerry's plate because he had a lot on it for 3.0-Alpha. I offered to take up this issue that wanted to use octokit/octokit instead of octokit/request.

I have spent a few hours looking into how I might go about this. Jerry suggested it would be relatively straightforward, but I am having trouble grasping what exactly needs to be done. Octokit shows up in a lot of different places in our repo

I need to have a meeting with Jerry to see if I am replacing all of the octokit packages with octokit/octokit or strictly just replacing the use of octokit/request for now.

That was the week that was for me this week. The final countdown is on now. 3.0 here we come!

Round the Corner

Tim Roberts — Sat, 02 Apr 2022 01:10:44 +0000

With there only being a few weeks left of this semester the end is in sight. We have one more release before our 3.0 release. It's time to start trying to wrap things up and get some final things landed and smoothed out, as opposed to adding a bunch of new features that will in turn add more bugs.

During our triage meeting this week I had made the suggestion that we cut some issues from the release tracker. This would help us prioritize what needs to get done for 3.0 and what is just a 'nice to have'. We managed to slim down the number of issues in the release by a substantial amount.

As it turns out, most of the issues I had left over are not really that important for 3.0. My focus for the coming weeks will be to help out where I can. I will be doing more reviews, and asking people if they need help.

The week that was

Even though it was not a necessity for 3.0, I decided to try and finish slimming down our registry images. I felt like I had spent so much time researching this that it would be a shame to have it all go down the drain and for someone else to have to start from scratch next semester.

After a bit more research in the topic however turned up this open issue in the docker-slim repo.

Turns out that docker-slim cannot be used as the building engine for buildx just quite yet. So this issue will have to be put in the backlog until it can be, because using buildx is more important than slimming the images.

Another issue that was left over to do now that the registry was up and running was to remove the old Github webhooks that we had in the repository. After confirming the location of them I simply clicked the red button and that was another issue taken care of.

Lastly, there was this issue that stemmed from migrating to use our registry based images. We need to update the gitpod.yaml to either build or cache-from like we now do in our development.yaml. I made this PR to address the issue and it is still a work in progress, but I think I am close to getting gitpod back up and running to help facilitate all this reviewing I plan on doing in the coming weeks.

Down With the Sickness

Tim Roberts — Mon, 28 Mar 2022 00:42:01 +0000

I am a little late this week with my blog. I usually write these things on Thursday, or Friday night. This week however I was hit hard by a sickness, late Thursday night. It is now Sunday night and I am finally feeling somewhat normal. Still a bit sore but the light sensitivity, fever, and headaches are all gone so I figured now would probably be a good time to get caught up on my work.

Release 2.9

In last weeks blog I talked about what my plan was for release 2.9. My main areas of concern was finishing the migration to make use of the images stored in our Docker registry. The other area I was planning on taking on was to slim down those images in the registry by using Docker-Slim.

I was able to submit a PR for migrating to use the docker based images. The only problem with the PR is that it is not testable, but with a couple reviews we took an educated guess at what needs to happen.

We to another educated guess with my second PR of the week. This one had me updating the autodeployment webhook to use our custom deploy tag that dynamically gets set depending on if it's staging or production. Again, this one is not testable in the local environment, so around 4-6 of us got together one night and tested it out on staging. After testing it out and smashing out some bugs, I made one final PR for the week that was just bringing the changes we made while testing live on staging, into the main branch.

Release 3.0-alpha

As you may have noticed, I did not get around to slimming the images like I had planned. This was because updating the autodeployment webhook was deemed to be a higher priority than slimming the images so I devoted my time to make sure that I could get that landed in time for 2.9.

Other than slimming the images, I believe there are still a few last steps needed before I can say that the registry is fully operational, but I am well on my way to being able to say I owned this registry.

The End Draws Near

Tim Roberts — Fri, 18 Mar 2022 00:54:44 +0000

This week's blog is going to be a little different than the others. Instead of doing a walk-through of the work I've done, this week we have been asked to plan for the inevitable conclusion of our course. The following questions have been given to us:

1) Which part of Telescope 3.0 do I own?
2) Which existing Issues are part of this? Which new Issues do I need to file in order to properly describe the work? When will those issues be filed?
3) Who can I depend on for support in development, debugging, testing, and reviews? I can't do this alone.
4) What are the risks that I see, which could prevent me from getting this shipped?
5) How will I mitigate these risks?

Which part of Telescope 3.0 do I own?

As mentioned in my original planning blog a few weeks back, I plan on owning the new Docker registry we have set up. Most people in the course are aware I'm the Docker guy, even though I know there's others in the class who probably have a better understanding of it all than I do 😅

Which existing Issues are part of this? Which new Issues do I need to file in order to properly describe the work? When will those issues be filed?

My main area of concern right now for release 2.9 is attempting to slim the Docker images. After going through the process outlined in issue #3189, to start actually using our registry images in staging an production, I must do the same process for all of the other microservices used in the project.

Once the images have been slimmed and our registry put to use, I think I can confidently say that "I own Docker", or at least the registry for Telescope.

Who can I depend on for support in development, debugging, testing, and reviews? I can't do this alone.

Kevan has been a huge help in getting the registry up an running. He was the one that did the majority of the authorization portion needed for the registry.

Josue, who is an alumni from the course, and who has also helped me out a great amount in setting up the registry, will be of great assistance on things to come, I am sure.

Finally, the professor himself, HumphD, will also be a key player in the success of this. As he has been for entirety of the course.

What are the risks that I see, which could prevent me from getting this shipped?

I seem to be hitting a rut here this week. I don't know if it's just the daylight savings time kicking my butt, if I'm fighting something off, or what is happening, but I have found it hard to concentrate on work/school this past week. Hopefully I can snap out of this daze soon and start feeling productive again.

Another risk is not asking for enough help. I feel like I am finally getting a better understanding of how you can configure and use a Dockerfile, and docker-compose, but I know I still have a lot to learn. I have done some more reading into the issues on deck and I feel like I'm at a loss for what to do. I just need to remember that I'm almost always lost and that my educated guesses have been spot on or pretty darn close so far, just gotta trust my gut and reason it out like I always end up doing.

How will I mitigate these risks?

In order to mitigate these risks, I just need to follow my own advise from last week and send up a PR so that I can start collaborating. Even though I am taking ownership for the registry, it's still a team effort.

No Rest for the Wicked

Tim Roberts — Sat, 12 Mar 2022 01:45:21 +0000

It's been a couple weeks since my last blog on account of last week being reading week. For me, it felt like any other weak because instead of going to class's I just did more work, and I still had to do school work as well so that I don't fall behind. Such is life.

Setting up the Registry

The moment has finally come. As I write this the final piece of the puzzle has just landed for setting up our docker image registry.

This has been where all of my attention has been going for this course over the past couple of weeks. I started it off by drafting up an initial version of how I thought the registry files should look.

After getting a bunch of input and making a similar amount of adjustments it was time to test it out. In order to do that I had to set up an ssh connection with the hard drive we are using for our physical storage.

I had some issues with that. I first gave the professor an ssh key that I had generated a wile back so that he could add me as a user. I tired to connect to the box using moba Xterm which is an ssh client i had been using for work, but for some reason it would not allow me to connect to the box. After trying that, I asked for some assistance and found that it was likely because the key was not being read.

I went back to ol reliable ssh client that I had been using for school on the recommendation of one of the best professors I had at Seneca Les Czegel (R.I.P 😢 ). After playing around with the configuration I thought I had it figured out for sure, but it still wouldn't let me connect. So I then generated another key, this time right from my ssh client's built in interface, and asked Dave to add that key as a user since I thought the other key was faulty (of course it's not my fault). Once he added that key as another user, I attempted to connect again...and it didn't work, again. I finally bit the bullet and downloaded PUTTY, and followed the instructions Dave had linked me in the picture above and it worked like a charm, with the original key that I had wanted to use.

Okay, great. Now that I was finally connected into the box it was time to fire up the registry. With a little help from Josue I was able to fire up the registry and push an image to it!

There was another couple of small changes to make before landing the initial files. There was also some follow up issues that needed to be created.

Teaming up on the Authorization

The main follow up issue that was created was the last piece of the puzzle needed in order to get the registry officially up and running.

Throughout this process Kevan has been interested in it and following along with it. He asked me if I had started on the auth stuff yet and I hadn't. He then asked a question in Slack that lead me to believe that he was looking into the issue.

I had a registry file in the first version of the registry files but did not land it because it was not being used at that point. the registry config was needed however for the auth process. As you can see in the screengrab, I had offered to send a PR up to get the ball rolling and to start the collaboration process. Luckily, before I had started the process Kevan had informed me that he already had something started, and then there was this little exchange on Slack which was nice to hear.

So after Kevan got his PR up, I then made a PR to his fork with the missing registry file and a couple of other minor adjustments. In the end, there where three of us that contributed to the one commit.

This was a great learning experience. I got to build a self hosted Docker registry from the ground up. I feel like I have a lot better understanding of how these dockerfiles and config files, work together and what a volume actually is. It also makes me sad that all of my professional option level courses will be done after this semester. I wish I had known Dave had a cloud computing course before I took the Azure course I am taking. He touches on all this, real world applicable, information in that course apparently.

Registering Images...Soon

Tim Roberts — Fri, 25 Feb 2022 23:40:13 +0000

This week was another busy week for me. I was trying to get ahead in my other courses in anticipation of moving all of our Docker images in into a registry that we will be setting up. I planned on taking on a lot of the tasks involved with this process, hence why I was a little light on the issues to work on this week. I had meant to do a couple of reviews to make up for this fact, but the time got away from me, and a couple of the other projects took longer than anticipated. But I have gotten all I need to get done for reading week for my other courses, so I can dedicate a lot of time to getting this registry set up over the next week.

PR for the week

I did manage to get one tiny PR landed this week. It was a follow up PR that I had mentioned last week. I just had to update the deployment file to mention the inclusion of the GITHUB_TOKEN

The time has come

Finally, the moment I've been waiting for. This issue had been on my list for a few weeks now, and we finally have the server in place and ready to set up as our registry.

I have done a lot of reading to get ready for this including:

I also gave my professor an ssh key so that I could make a secure connection with the private registry we are setting up.

Reading Week

This coming week is reading week at school. As mentioned before I will be devoting a lot of my non-work time to getting this registry up and running. I feel like this task is a bit beyond my current level, but there's only one way to learn these things, and that's by doing it.

Coding in the Dark for 2.7

Tim Roberts — Thu, 17 Feb 2022 16:56:22 +0000

This was a busy week for me. I managed to get three PR's submitted all while having some extra meetings and deadlines show up for work. Not to mention getting this little two month old cutie on Saturday (Feb 12th) and dealing with everything that having a puppy entails.

Adding GITHUB_TOKEN to the environment

The first PR was actually one I submitted last week as a draft because I needed some help figuring out why my approach wasn't working.

As you can see from the description in the PR, my wording shows I have my doubts about if it will work or not. I was really just taking educated guesses on everything and most of my assumptions ended up being correct. I only had to make a few minor adjustments and it was time to test.

I had no clue where to even begin testing to see if this was all working. Josue helped out a bunch with this/ he did all the testing while I watched in awe through screen-share. I knew he was a core contributor and an integral part of Telescope, and I got a glimpse as to just how important he is to the project.

In order to test it out Josue had to shut down the staging environment. Then configured the files to be like what I would be submitting in the PR. He had to do this because the PR would potentially delete things and not build them back properly if not working correctly. He then had to go into the settings of the Telescope repo and trigger the web hook to mimic a PR getting merged. That confirmed the GITHUB_TOKEN was indeed populating.

This PR was a very informative process. Not just for me either.

Reducing the size of web service

The second PR was another one I started last week and had made a draft PR for because I needed some assistance. Once again it was Josue to the rescue.

After making those fixes I got the green light from the professor.

Just waiting on at least one more review before landing this one.

Dockerizing Docusaurus

The previous PR's I went over above either had no release attached to them, or was scheduled for 2.8. The only task I had for 2.7 was this one. Unfortunately I missed it being assigned to me last week and only noticed it on Monday. Started working on it Tuesday, and submitted yet another draft PR because of roadblocks I was hitting.

After submitting the PR, I noticed Cindy had started working on it a few hours later. She came across an issue when building it that she filed an issue for, then submitted a PR for. I reviewed the PR to help facilitate the assistance she was providing.

After coming back to the issue the next night, it would seem Cindy and Dave have got it sorted out!

Coming up next

There has been some issues that stemmed from some of the work I have done this week. I need to dig around the repo and find any Dockerfiles that I could slim down by doing what I did for my second PR of this week. I can then file issues and start submitting PR's for each one.

I also need to submit a PR in follow up of my first PR of the week.

I had meant to actually make that simple change to the doc and amend the commit for that PR, but I was too eager to get it landed and start working on the Docusaurus issue.

A Glimpse into my Future

Tim Roberts — Tue, 08 Feb 2022 22:33:54 +0000

Welcome to my first blog on dev.to! It would seem that most of you have not likely read any of my blog posts yet because there has been an issue with my old WordPress blog not getting picked up by Telescope. For those of you living under a rock, here is a little reminder.

Usually, we write a blog about what we have done throughout the week. This week, the professor would like us to write about what we are currently doing, or planning on working on for the week.

First Order of Business

I have already been looking into some other things, but I will put those on hold for now in order to complete this week's blog.

A lot of hours have been spent by the team debugging the issue. At first it seemed like a 305 error, but in testing it was not easily repeatable. Sometimes a 305 would pop up, other times it was a 500 error, and then other times it would give no error. The professor then thought it might be a cashe issue, so after getting Josue to flush the cashe, my feed was still not getting picked up.

The professor finally found out that my feed appeared to be empty. I did a little more research into the issue. Apparently there is supposed to be a feed: url inside the meta object of my blog. After a some time searching for it I could not seem to find it.

I figured since I started that blog over two years ago, and left it inactive after finishing the previous course could have been causing the issue. It was also only my blog that was causing an issue, and I was the only person using WordPress for my blog. I decided to join all the cool kids and start a blog on dev.to.

Current Issues

The issues I would most like to work on are the ones that are on hold. For 'Making Telescope usable by registry' We are now waiting on the confirmation from Docker to see if we qualify for there special program that would allow Telescope to use DockerHub for free without restrictions.

Slimming our images is dependent on our images getting registered so once we figure this out, I will have a lot of work to do. In the meantime.

Issue 2835 Seemed like it would be an easy one to start with this week. I first checked with Francesco to see if he started working on it, and he had so I moved onto the next issue.

Turned out that Francesco has ran into an issue for it and would like some assistance. I have already taken a glance and came up short, but will spend some time with it later.

Before Francesco asked for help on that issue, I had started looking into issue 2799 about using the GITHUB_TOKEN in all of our env files. I have been poking around the autodeployment folder for a bit based on Josue's recommendation

I think I will have to edit the lines found here in the deploy.sh file

I'm not certain that that is the right spot to alter the code, so I am going to spend a little bit more time looking at it, and if nothing clicks I will then ask for some clarification. I'm also not really sure how to go about testing that it works. A theme that seems to be coming up often.

Wrapping up

Well there you have it. That is my plan of action for this week in Telescope. As you can see from my issues, I am really trying to stick with my goal for 3.0 and do all things Docker. Hopefully my feed gets picked up this time so I can stop having my name show up in issues.