Results of Slurm-3

On Sunday, Slurm-3 ended: intensive on Kubernetes.

We set a personal best: 132 participants, 65 online and 67 in the hall.
On the first Slurm there were 50 people, on the second 87. We are growing slowly.

Results of Slurm-3

126 people created a cluster (the task of the first day), and 115 completed the practice to the end.
6 people ignored the practice altogether. Let's assume that they needed only lectures.

The main fakap this time was related to the broadcast: either the speaker's microphone was cut down, or they forgot to turn off the music. It's time to move from the collective-farm method "professionals will figure it out" to regulations.

Two more problems were related to Git. Firstly, the participants found task materials there and rushed to perform ahead of the locomotive. As a result, we reach practice, and the person has everything broken there. Next time we will push materials as needed.

We also made files with commands, because last time people copied commands from pdf along with formatting characters, and nothing worked for them. Having discovered these files, some cheerfully rushed to copy them to the console, and then complain that the practice was reduced to copy-paste. The practice was reduced to completing tasks, configs and commands should have been written manually, no one forced to copy-paste.

Results of Slurm-3

Feedback was given by 46 people. We will assume that this is a representative section of the audience.

How do you like the intensity of Slurm?

33: just right.
10: too simple and slow, would like more material
3: too fast and complicated, I would like less material.

We normally fall into the declared audience: those who get acquainted with Kubernetes.
For those who find the usual Slurm too simple, MegaSlurm will take place in early June. We give a discount of 15 thousand to all participants of the basic Slurm, Slurm-3 will pay off anyway.

Has Kubernetes become clear?

16: I knew k8s before, now I know better.
13: k8s did not know before, now I figured it out.
15: I don’t understand k8s yet, but I see where to dig.
2: I did not learn anything new.
0: I did not understand anything about k8s.

The situation is even better than I imagined. I thought that at least half would answer “I don’t understand k8s yet, but I see where to dig”, and there will be practically no those who “k8s didn’t know, now I figured it out”.

Results of Slurm-3

The interesting starts next. 6 people did not solve the problem with which they went to Slurm. Four had specific requirements, we will take them into account when developing the MegaSlurma program. And here are two reviews in full (with minimal editing):

Monotonous storytelling with a lot of water
Illiterate speech with a bunch of jargon
A huge amount of time is devoted to third-rate trifles (half an hour we change numbers in configs? intensive, yes)
Irrelevant presentations
ceph is not needed
The approach of a skeleton builder is visible in everything

This slurm is not for beginners and not for experienced.
Irritated by both slowness and speed, and as little information as possible:

  1. For some reason, the presenters rested on details, on variables (indistinctly explaining why they are needed).
  2. They quickly ran through practice: “here it is ... hop-hop-hop works.”
  3. In theory, only Pavel has the fewest questions, to the rest of the speakers: why is it so dull and uninteresting and I want you to finish it faster? Still nothing is clear.
    At some point you want to do this: What??? What was it now??? Why is this all? >Why, without explaining, run on??? It doesn’t work for someone, but the leader flies on ... WAIT!!!
    As a result, I want to get up and leave, but 25 tr. put you back.

I feel bad that these members didn't text me on the first day. There was a dissatisfied person on the second Slurm, he called me, asked for a refund, and we immediately disabled his access and returned the money.

For the next Slurm, I will prepare the return rules so as not to torment those who Slurm did not visit.

But in general, 2 negative reviews per 46 respondents (and 132 participants) is close to ideal.

Results of Slurm-3

Finally. A member of Slurm-1 recently wrote to me that he is still reviewing the records and finding something new in them. So graduating from Slurm doesn't mean graduating.

Thanks to everyone who was on Slurm!

Anton Skobin

Source: habr.com

Add a comment