Live data from Hacker News

Not only coders are hard to recruit

blog.sysadminsarena.com

21–30 of 73 posts

Re: Not only coders are hard to recruit

#21

Earlier quoted context omitted.

I don't think I would enjoy HackerRank as an interview, but as a user I do enjoy being able to practice computer science problems.

Just wrote an answer above which i think is good for your concerns. What'do you think?

I think if that's the approach SysAdminsArena is taking, it's quite good. More advanced problems could be replication, caching, etc.

Re: Not only coders are hard to recruit

#22
post #14

Earlier quoted context omitted.

Just wrote an answer above which i think is good for your concerns. What'do you think?

(Speaking as a coder here, although I've also done sysadmin through most of my 20 years of working in the industry). This would be quite nice. Going through the job searching process at the moment, the "aptitude tests" etc that I've seen are very good at blowing cobwebs off my obscure analytical skills, which I rarely if ever use, yet fail completely at getting any gauge of my ability to get every day stuff done in a…

You just mentioned some of reasons why we are making this product. If you have more ideas, please share with us. Thanks

Re: Not only coders are hard to recruit

#23

So this is to have a test VM they have to fix? Excellent! I heartily approve of this method for finding a sysadmin. This is a thread where I talk about how we did this: https://news.ycombinator.com/item?id=8332104 We used to use it a lot - a trivial Java webapp running in a slightly misconfigured Tomcat with a slightly misconfigured Apache in front on a slightly misconfigured Unix box - though it's fallen by the ways…

Hi David, thanks for your thoughts. Very useful the thread provided. And YES, this is what we are doing. Keystrokes is just one of the metrics we are about to collect and analyze. We are also thinking about number of commands wrote, how many were correct, how many wrong, also a log of the entire test. In future phases, we may take into consideration pair working or capturing a movie of the test. What other ideas of r…

So, if I will follow best practices and will prepare backup, then restore backup locally (to check that backup is OK), then will play with backup until problem will be solved (to not harm target server), then prepare runbook, and then will just execute it at target server, I will pass your test? Just two commands.

Re: Not only coders are hard to recruit

#24
The Cisco Certified Internetwork Engineer (CCIE) certification used to be the gold standard amongst network engineers 10-20 years ago.

Becoming a CCIE required a passing a notoriously difficult hands-on 8 hour in-person lab examination[1]. SysAdminsArena reminds me of a remote version of the CCIE Lab Exam for sysadmins.

Fun fact: the CCIE program was established in 1992 and was originally going to be called "Cisco Top Gun"[2], but they had to choose a more serious name.

[1] https://learningnetwork.cisco.com/community/certifications/c...

[2] http://www.bradreese.com/blog/how-the-cisco-ccie-program-was...

Re: Not only coders are hard to recruit

#25

Earlier quoted context omitted.

Please allow me my honesty but that's a stupid way of thinking about candidates and the roles. Or I do not get it. So you are basically going to judge the candidate based on what criteria ? If they use vi instead of cat to append to a file ? So decide what is the best way of solving the problem ? Or you give +1pt for using tail instead of opening the whole file in VI ? Are you aware of the fact that there are many ty…

We are aware that there are more ways to skin a cat. I think it's fair to evaluate a candidate based on wether if he can solve an issue or not and the amount of time it took. I agree it would be stupid to score based on individual commands and we didn't plan to go that low level. But we plan to count the number of commands with typos or wrong parameters. Don't you agree this would be a way to differentiate between 2…

Sorry I have to disagree here as well. I do not think the time or number of commands or typos should be measured. I do not think you can do any quantification here. I believe a lot of sysadmins (talking about unix) do w;uptime;inspect syslog as soon as they connect to the system. Most of the time it's just not necessary to do this but for a lot of sysadmins it's a habit that actually saved their asses several times.

Does it make sense for a sysadmin to know a numeric notation for chmod ? Probably. Would I ask that on an interview ? Nope If you need to ask why then you have probably never worked in a heterogeneous environment.

Re: Not only coders are hard to recruit

#26

Earlier quoted context omitted.

Glad you asked that. As i saw in the treads mentioned by you, the common problem is that the tests are more about theory, and less about practice. Which is true. Sometimes, in interviews, you are asked about algorithms that you studied in college and never used before. It's very difficult to score high in this tests. We propose to eliminate this gap by having real job, day to day, scenarios like: install apache, mysq…

Those are pretty baseline things that get automated away pretty quickly. More useful to me (as both a practicing devops type and a manager in that capacity) would be tests gauging ability to write bash scripts to do various things, or simulated problems that need to be diagnosed/fixed using command line tools (e.g., strace, netstat, etc.).

Indeed, those are pretty basic tasks. We will have 3 levels of difficulty and the examples mentioned are in the easiest category. In a early phase of a recruitment process, an experienced sysadmin would solve in less than 3 minutes and this is helping screening process when you have a huge number of applicants. But for latest stages of recruitment process, we will have more advanced scenarios for the tests.

Definitely your suggestions are great and we will discuss having tests in this area.

Cheers!

Re: Not only coders are hard to recruit

#27
The premise that technical people with in-demand skills are hard to recruit has very little to do with the screening process. The screening process is not a difficult problem to solve.

The real difficulty in recruiting in this space is finding and engaging these people, not testing them.

If this is a Hacker Rank for sysadmins, then great. I'm sure you'll have plenty of users but we're still a long way away from solving the bigger problem.

Re: Not only coders are hard to recruit

#28
Interesting problem you're addressing but I've been working on this problem from a completely different set of metrics. Fundamentally, I really don't care how a problem is fixed as long as it doesn't have bad repercussions. If a web server is "slow" users don't care why - giving freedom to admins to approach higher level problems while still measuring them is something I'm confident that can be done while I take a number of issues with assessing anything based upon commands entered even in realtime (I'm mostly concerned from a security / bad or illegal practices in prod perspective, to be frank).

The approach I'm working on is easier to do technically but requires some more costs.

With that said, glad to see I'm not the only one that's been frustrated with hiring devops / sysadmin / SRE roles. Google's even commented how hard hiring for SREs is for them and I think we have a huge gap between people stuck in the 90s approaching systems and those that modernized.

Re: Not only coders are hard to recruit

#30
Isn't everyone hard to recruit? The difference in perception of coders might be that the evidence for a bad recruit is obvious in the code they produce. There could be many bad recruits in all other areas but are less obvious.

I generally like the idea of clearer quantification of a persons output but it should be done very delicately. As in, if you are going to measure everything someone produces, maybe that data should only be available to that person and it's up to them how they share it.

The uncertainty principle likely applies to people too. Hard to measure people without influencing them. Trick is to do it in a positive way. Managers measuring people for the purpose of firing them would be perceived, and have a negative effect on morale, than measuring people to identify opportunities for education, training or skills development.

Post reply on HN