83
submitted 2 months ago* (last edited 1 month ago) by thelastaxolotl@hexbear.net to c/chapotraphouse@hexbear.net

The HESA Shahed 136 (Persian: شاهد ۱۳۶, lit. 'Witness 136'), also known by its Russian designation Geran-2 (Russian: Герань-2, lit. 'Geranium-2'), is an Iranian-designed one-way attack drone, also referred to as a kamikaze drone or suicide drone, in the form of an autonomous pusher-propelled drone. It is designed and manufactured by the Iranian state-owned corporation HESA in association with Shahed Aviation Industries.

The munition is designed to attack ground targets from a distance. The drone is typically fired in multiples from a launch rack. The first public footage of the drone was released in December 2021. Russia has made much use of the Shahed 136/Geran-2 in the Russo-Ukrainian war, especially in strikes against Ukrainian infrastructure, and mass-produces its own version.

Description

The aircraft has a cropped delta-wing shape, with a central fuselage blending into the wings, which have vertical stabilizing rudders at the tips. The nose section contains a warhead estimated to weigh 30–50 kilograms (66–110 lb). An Iranian-made Mado MD-550 engine sits in the rear of the fuselage and drives a two-bladed pusher propeller. The MD-550 is reverse engineered from the Limbach L550E, a 550cc four-cylinder two-stroke petrol aircraft engine made in Germany. The munition is 3.5 metres (11 ft) long, with a wingspan of 2.5 metres (8.2 ft), flies at over 185 kilometres per hour (115 mph), and weighs about 200 kilograms (440 lb).[16] The drone's appearance resembles that of the Drohne-Anti-Radar (DAR) developed by Dornier Flugzeugwerke in Germany in the 1980s, but whether there was actual copying is an open question.

Its range has been estimated to be anywhere from between 970–1,500 km (600–930 mi) to as much as 2,000–2,500 km (1,200–1,600 mi). The U.S. Army unclassified worldwide equipment guide states that the Shahed 136 design supports an aerial reconnaissance option, although no cameras were noted in the Geran-2 in Russian service.

Electronics

The Shahed 136 navigates by a commercial-grade inertial guidance system, corrected by civilian GPS and GLONASS. December 2023 remains from the drones were found with SIMs and 4G modems of the type used in mobile phones.

Deployment

Because of the portability of the launch frame and drone assembly, the entire unit can be mounted on the back of any military or commercial truck. The aircraft is launched on rails at a slight upward angle with initial rocket launch assistance. The rocket is jettisoned immediately after launch, whereupon the drone's conventional piston engine takes over

Loitering munition

A loitering munition (LM) is a type of self-destructive unmanned aerial vehicle (UAV) equipped with a warhead that is typically designed to remotely loiter by a human operator using electro-optical targeting sensors or camera suite and data-link until a target is designated, then crash into it and detonate. Anti-radiation (anti-radar) loitering munitions are a type of loitering munition that employ either an anti-radiation seeker by itself or in tandem with an electro-optical targeting system to locate enemy radar by loitering and destroy it after detection. Common terms like suicide drone, kamikaze drone, or exploding drone are used for both loitering munition and one-way attack drones. They enable attacks against hidden targets that emerge for short periods without placing high-value platforms near the target area. Unlike many other types of munitions, their attacks can be changed mid-mission or aborted. Loitering munitions are typically aerial platforms, but include some autonomous undersea vehicles with similar characteristics.

Loitering weapons emerged in the 1980s for the suppression of enemy air defenses (SEAD) role, and were deployed for SEAD by some military forces in the 1990s. In the 2000s, they were developed for additional roles, from long-range strikes and fire support to short-range tactical systems that fit in a backpack.

Megathreads and spaces to hang out:

reminders:

  • 💚 You nerds can join specific comms to see posts about all sorts of topics
  • 💙 Hexbear’s algorithm prioritizes comments over upbears
  • 💜 Sorting by new you nerd
  • 🐶 Join the unofficial Hexbear-adjacent Mastodon instance toots.matapacos.dog

Links To Resources (Aid and Theory):

Aid:

Theory:

Financial Support to the Bearsite

you are viewing a single comment's thread
view the rest of the comments
[-] Chana@hexbear.net 8 points 2 months ago

Back to debian 😔.

GUIX was cool but really not right (yet) for either desktop or server use, at least for me. I really like the thing I was looking for in it, which is declarative configuration of nearly the entire system and that part worked flawlessly. And scheme is fun. But here are some significant pain points:

  • They only have Linux Libre in the main repos. Linux Libre is a pointless security hole whose only practical purpose is to nudge people to get hardware that uses open microcode or that has vendors pushing bios updates for as long as you have your device. This is basically zero personal computers. It is so thoroughly patched that you can't even feasibly load your own microcode on top of it with any reliability. So to have basic security with a normal computer, you have to use the repos and kernel of the undocumented sister project nonguix.
  • Nonguix being undocumented wasn't a huge issue for me, but having something as critical as the kernel have such little focus and upkeep is a disaster. Just trying to pin the kernel to an lts release was surprisingly challenging due to the dual repos needed and them being out of sync re: 6.12 and 6.18.
  • The nonguix repos were effectively down for like a week for "substitutes" (binaries), meaning you'd have to either wait (like me) or compile the kernel etc, which takes quite some time.
  • The main repos broke my browser for over 3 weeks and so I switched to flatpaks. I also realized it was using a version from over 6 months ago, largely unlatched. My browser was broken at the packaging level, so I could not update my system (or at least half of it) until I decided to partially abandon the main draw for me, declarative management.
  • The guile parts were oddly slow. guix shell is cool but using it to make a system container that stimulates a typical Linux filesystem has a 1-5 second start time due to mostly guile itself taking a long time to load modules. But this is also the wrapper tool! So if you want to use a snappy cli tool that needs this emulation, you either now have a clunky one or you have to (like me) do things like write a bubble wrap script with no fewer than 30 flags and yes, their exact order is very important.
  • No real "offline reconfigure". You can tell it to not use substitutes and it will work that way, but only if nothing needs to be downloaded, so you have to sit there and watch it to be sure or else it might download compile something. This means every reconfigure has a 5-20 second delay while it gathers its guile modules and checks for binaries remotely that it won't need to download. Then it reconfigures to add my one setting change in like a half second.
  • Various applications are much slower for whatever reason, even with flatpaks. Possibly due to the nonstandard filesystem and need to do elf patching, but I don't really know.
  • Modularizing your config requires creating a guile module. That part is easy and fun but now you have to either add it to your guile paths for the things you define to be recognized by your reconfiguration commands and ide you use to edit them. Or turn it into a "channel". If you add it to your paths, it's advised against doing so globally via your shell environment. This leaves you relying on temporary environment variable declarations and inconsistent cli tool flags. And adds oddly large amounts of load time overhead. If you forget to add just the right guile module path flags to just the right programs, you will get an opaque error, usually in the parent calling module, and have to guess and check. If you use a channel, you can't actually use any changes you make until you commit to the repo, so you can't do a "add this, run, then commit" workflow. You have to switch up "add this, commit, run, and revert if it broke" workflow, which I hate.

Really the eventual deal breakers for me were the security issues (linux-libre, nonguix linux servers down, unupdatable system, old and unpatched security critical packages) and the necessity and difficulty of FHS filesystem emulation for tools I needed, but it all also just kind of added up until I had to give up on it because I couldn't practically work around some of these limitations, namely the kernel and packaging issues. The former being a poor design choice and the latter being an issue of design choice (rolling distribution) and not having enough maintainers. I'd have started helping with the latter issue myself but I wasn't sure I would be sticking around.

Anyways it is still a very cool system and is not itself a bad thing at all despite my criticisms. The actual core system is very cool and structurally sound aside from the fact that it should probably use fewer macros (not lintable) and guile needs a faster engine. I hope GUIX does well and I hope I can try it out again with some of these issues resolved and start making my own contributions. But back to the grill pill FOSS distro, debian my dearest.

[-] ajab@hexbear.net 2 points 2 months ago

So are you going to use it on Debian or is it joever for GUIX?

[-] Chana@hexbear.net 2 points 2 months ago

Joever. I don't really get anything out of it if I'm not being declarative with 90% of the system. My system is already like 80% declarative with just a package list, defining 8 configs, and saying what services should start. That last 20% is stuff that would require me to make it my system OS again.

It's a bummer, man.

this post was submitted on 20 Mar 2026
83 points (100.0% liked)

Chapotraphouse

14372 readers
784 users here now

Banned? DM Wmill to appeal.

No anti-nautilism posts. See: Eco-fascism Primer

Slop posts go in c/slop. Don't post low-hanging fruit here.

founded 5 years ago
MODERATORS