v1.0
Linux World Cup (LWC) first edition: 2023 is done, see ranking
LWC is a Linux-related online competition.
A lot of people enjoy Linux challenges and it can be fun. There are coding competitions, why not a Linux one?
Open to anyone with Internet and a browser.
The "Qualifier" for the first edition was held Sunday September 24, 2023.
The "Finals" for the first edition was Sunday, October 1st 2023.
SadServers.com will be the platform that manages the challenges. See the "Mechanics" section below for details.
It's free. (I'll be happy if you "buy me coffee" to help with hosting costs).
There are no guaranteed material prizes.
Next: announcing the Second edition of the Linux World Cup for 2024.
-
Participants must come from a "regular" IP address that can be resolved geographically, that means for example not using a VPN provider, and not coming from a cloud provider server. I consider IP address/geolocation as private information. Contact me if this rule is a problem for you.
-
There's a deadline before the competition starts (see Timeline section below) to change your timezone. Participants must be in their declared timezone and use the corresponding competition times for it.
-
Please do not delete or try to delete your shell command history. Command history is not private information and may be publicly disclosed.
-
Competitors can use any information at their disposal, including AI tools like ChatGPT and similar, as long as it is individual work (other people helping is not allowed).
In general the scenarios and VMs will behave the same as any other SadServer scenario and VM with some changes to help running the competition smoothly:
- VMs won't be AWS spot instances but regular on-demand instances.
- Rate limiting is disabled for the event participant.
Also:
- There may be no "clues" provided for competition scenarios (there may be tips but they would be in the description, not under the "Next Clue / Solution" button).
- Scenarios will tend not to have root (sudo) access.
(This may change in future editions).
Competition will be done in two different Sundays:
- One day for a "qualifier" event (one scenario). No ranking or cut-off, it's pass/fail.
- Another "finals" day for those who pass the qualifier phase, consisting of several scenarios. The final scenarios can be done in any order (ie you don't have to pass one to access another). I'll order them by (my perceived) difficulty: easier to harder.
During both those days there will be three four-hour windows of time to participate, one interval per each of the three time zone groups: Asia-Pacific (UTC+4 to UTC+12), Africa-Europe (UTC-3 to UTC+4), Americas (UTC-11 to UTC-3).
People on all time zones will compete on the same date on a Sunday. People in some locales may find more convenient to use a different time zone area than the continet suggested.
Please remember you can only use one time zone area and it must be declared in your SadServers LWC account page.
Sunday from 00:00 UTC to 04:00 UTC
Asia-Pacific
Examples of local times:
New Dheli 5:30 am - 9:30 am
Beijing: 8 am - 12 noon
Tokyo : 9 am - 1 pm
Sydney: 10 am - 2 pm
Sunday from 15:00 UTC to 19:00 UTC
Europe-Africa
Examples of local times:
Dakar: 3 pm - 7 pm
Berlin: 5 pm - 9 pm
Kyiv: 6 pm - 10 pm
Sunday from 20:00 UTC to 00:00 UTC
Americas
Examples of local times:
Rio: 5 pm - 9 pm
New York: 4 pm - 8 pm
Vancouver: 1 pm - 5 pm
Honolulu: 10 am - 2 pm
Participants can try the challenges at any time during their window; there's no advantage in doing them starting at an earlier time (it may be disadvantageous to try right when the window opens if there's many people doing it at the same time).
(This may change in future editions).
I wasn't sure how to do or organize teams when I created the sign up form; if maybe people were interested in pairing up with other people/teams they didn't know. Also I forgot to put a limit to the number of people in a team, and now we may have some large teams which doesn't seem totally fair. Anyways, my bad, for this event I simplified everything and at the end people are either in a team they put in the form or not. If there's only one person with a team name, that's fine, it's an individual participant.
In SadServers you'll be able to edit the team name or delete it if you want to change your status. There's currently no internal software logic regarding teams other than the team name associated with a person's account.
-
There are two tracks: individual and teams (ie one ranking for individuals and one for teams).
-
Scoring criteria: ranks will be created according to least total time for people/teams in solving all the final scenarios. If one or more scenarios are not solved (not expecting this but just in case), then ranking will be by least time in the solved scenarios (ie, final scenarios have the same scoring weight although they may have different difficulty).
-
The automatic check validation script in the readable ~/agent/check.sh is what decides if a solution is valid or not. In extreme circumstances I'll check the command history manually to validate a solution that didn't pass the check script.
The challenges or scenarios run on single, bootable VMs, so that restricts the type of issues to present.
The hard scenarios in SadServers.com can give an idea of how the LWC can be.
Not sure how useful this classification is but scenarios can be of the "fix" type (something is broken, troubleshoot and leave in a working state), or the "hack"/"capture the flag" type to uncover a secret or puzzle, or the "do/find" type, like manipulating data for example.
I favor the Debian distribution but other flavors of Linux that run on AWS can be used.
In some scenarios the default user can elevate privileges to superadmin (root), in some scenarios this is disabled.
The tooling (packages) installed in each scenario may be different.
Some programming skills (using a shell script or a popular programming language like Python) are required. Some basic SQL can be needed.
It's not possible to list the Linux components or tools that can show up in the challenges but popular Open Source implementations of key services like web servers, proxy servers, databases (relational or not) and message brokers are fair game.
Containerization (Docker) is an important technology that may be included. Kubernetes is out of scope for this competition.
I really want to run a nice and fair event.
-
Be Kind. AKA don't be a jerk towards other participants or organizers.
-
Don't Cheat. Hard to define all the possibilities but as rules:
- One person per account (registered email in SadServers).
- If you are competing in a team with other people, all members need to have their team name in their SadServers LWC profile. Otherwise I cannot correlate users and teams. Only one member of the team needs to solve the scenario; there's no need for the other members to enter a solution.
- Related to the above, scenarios where the solution is entered directly or out of nowhere without previous steps to support it, will be void (invalid).
- Do not share any information about the scenarios until after the event is over in all timezones.
- Do not try to manipulate or bypass the services working in the VMs that are needed for the overall functioning of the system and which are not related to the challenge to solve, like the solution checking agent or a command history logging agent.
TL;DR "Expect nothing" :-)
LWC and SadServers is run (and financed!) by just this guy (well, I've had help in creating some of the scenarios). I'll try my best but I can always make mistakes (there's a couple issues already in SadSevers scenarios for example). I may have forgotten an "obvious" rule or explanation here and they may change before the competition.
There's also the technical things that can go wrong, like network delays or issues in AWS.
A special pain point is the "Check Solution" script, which may not be perfect and have false positives and false negatives. I may review logs and decide to override the script output and yes this can be subjective.
While the event is running, if there's any system-wide issue or any important notification, I'll use the SadServers Twitter (erm, X) account. I may not be available at times during the competition; in principle it should run automatically without any manual intervention.
The SLA or QoS of this is "best effort".
Open an Issue here if you think it's useful for other people or contact me (info@sadservers.com for ex) if it's more individual.
Hi,
I'd like some clarification on the finals for the LWC.
I noticed on github that you mentioned "only one member of the team needs to solve the scenario".
Will there be separate scenarios for both the individual and team events? If not, how would the individual events work if others in the team have already solved it? How will you take into account the fact that teams can be in multiple time zones?
Hello,
Thanks for asking, I may clarify further in the Github repo.
I assume teams are working together as a unit; they can tell each other what they've found working in their scenarios or they may be doing the scenario together (screensharing or in the same room). I assume all information is shared at any time in full collaboration. The mechanisms don't matter and where they are or what people on the team are doing are not relevant.
The only thing that matters is the shortest time to first fully complete the scenario (shortest time in a hard scenario may include two sessions for example). One team member entering the solution as a shortcut when somebody else on the team has solved it doesn't count.
To respond more specifically, the scenarios are the same for both teams and members. As said in the repo, there are two "tracks" or separate competitions; one for individuals and one for teams, same as running (swimming/rowing) individual 400 meters or a team of 4 running 4x100 meters; same distance but teams win as one and all members are recognized.
I did not take into account that members of the same team can be in different time zones :-( but it seems to me that the best tor simplest thing to do in the future is to not allow it. It would be like running the 4x100 event but with 4 separate 100m races and pooling the results, it kind of defeats the "team" idea of working in collaboration. In our current situation if that happens, I'll allow it but will still go with the "shortest time to first fully complete the scenario" metric.
Let me know if there's still ambiguity.
Stepping back, with only one pair of eyes, it's difficult to figure out and predict edge cases and the way to go is to learn from mistakes for next editions of the competition and to have people participating as volunteers to help.
Thanks, Fernando