Skip to main content
A few random things that I have been dealing with

My Digitalocean droplet suffered recently a brute force ssh attack. Unfortunately I noticed it a couple of days after the attack had happened but luckily the act caused little harm except very high CPU usage from sshd process for a few days. I'm not sure how to really protect against such attacks (cheaply) but I decided to try out fail2ban. With fail2ban I could protect the server also against attacks towards nginx.

Installing it was simple enough and I saw it was working rather well. There were some 3K ssh login attempts per day and the iptables based port blocking reduced the amount to some hundreds. After a while though I noticed that the fail2ban stopped blocking unauthorized IPs. I took a look at fail2ban github and saw some issues with ssh regex filters (fail2ban works by monitoring logs and matching those against predefined regexes). I made some small adjustments but still no luck, it did not ban anything.

I turned on debug logs and saw lines where it said something like: IGNORE timestamp < another timestamp - findtime. I converted the timestamps to human readable format and checked those against auth.log. Those looked weird, the timestamps were way off. Then it hit me, i had recently changed the timezone of the server to the Finnish one. By quick googling around I found out that after the timestamp is updated, the syslog daemon needs to be restarted as well so the system log timestamps would match the system timezone. service rsyslog restart did the trick.


Another totally unrelated thing on the very same droplet was an issue with running a scala application as a systemd service. I ran a development version of the process simply with nohup previously but since the app was about to go to "production", i thought it would need something to keep it always up and running. I was somewhat familiar to upstart and it turned out that there were some good instructions which upstart keywords would match the systemd service files.

Writing the service file was a breeze, only a couple of lines. The problems started when I started the service. It was running fine for a couple of seconds, it printed out nicely the first few log lines to the standard output. However after the app was up (rest service was running) the service exited with success code 0. I tried modifying some of the basic configs like Type=simple and SuccessExitStatus=143 and even using sh -c to start the java process. After a while I was convinced that the problem was not in the service file but rather in the scala application startup. And indeed it was. After the rest service has started, I added a line StdIn.readLine() followed by a graceful stopping of the process. The idea is to hit enter to stop the application if it running on foreground. After I removed that line and the graceful shutdown, it was running fine as a systemd service. It seems systemd sends something to the process stdin when it is starting up / started.

tldr: Run service rsyslog restart after you have changed ubuntu timezone. Check std in reads on processes running as a systemd service.

Comments

Popular posts from this blog

Bird is causing high CPU on my macOS

There is no lack of people complaining about MacOS Tahoe, mostly about rounded corners and inconsistent design decisions. I have not paid that much attentention to that, but there is one mac bug i have paid attention to. It hasn't been a visual or ux but rather few system processes pegging the CPU. trustd , alongside with ecosystemd and ecosystemanalyticsd all reported high CPU usage. I can't exactly recall when this started, it might have predated my Tahoe upgrade but anyway, the trio of processes all had high CPU usage. Sure, it may have been the virtual efficiency cores and whether it affeced battery life or slowed down other processes i don't know to be hon...

How my work has changed in the past few years

When I started at my current job 3ish years ago, we had nothing but a few PowerPoint slides and a Figma prototype. I recall talking with one of the founders, whom I had worked with some 10 years ago, and pointing out that the very next step after PowerPoint and Figma was still the same as it had always been: translating those slides and mockups into CRUDs, forms, and other basic building blocks. Sure, all the new fancy tools and frameworks (we had jQuery then) help, but I still had to tinker with very basic primitives like tables and forms. I could not just say, "here is a table, here is a form." Well, yes, technically I could, but I still needed to write a lot of code to make a working form or table. This was around the time of GPT-3, Copilot beta, and very early Cursor. I did attempt to use them, but they could only perform very basic tasks and were pretty much useless for frontend work. We, being one of the first-wave "ChatGPT wrapper" startups has naturally pus...

Building a OTel sink that follows GenAI semantic conventions

I've been working on an OTel sink recently. To be more specific, it concerns receiving telemetry data from LLM interactions. Fortunately, there is a spec for all of this called GenAI Semantic Conventions . That is what I based my work on, so I can make assumptions about key attributes to highlight in the sink's UI, which is actually a CLI in my case. I dogfooded my handrolled sink using Pydantic AI (my favorite LLM framework), and all good, it follows the spec well. Key attributes flow in as specified. Pydantic AI also includes some of its own attributes, but that is expected. However, pick something like Claude Code, (via anthropic agent sdk), the reality is a bit different. It does not follow the spec at all. It has its own custom attribute layout . Same direction, different flavor with LangChain. The Python source has zero opentelemetry imports anywhere. None. The ls_* namespace ( ls_provider , ls_model_name , ls_temperature ) is its own thing, with LangSmith as the on...