Boffins find COVID changed the way sysadmins work – probably for the worse Researchers from Germany's Max-Planck-Institut für Informatik have studied how the COVID-19 pandemic impacted the way systems administrators work, and found the profession was negatively impacted. Their findings are detailed in a paper titled "'I needed to solve their overwhelmness': How system administration work was affected …
So much wrong with this study. Like you say 24 samples, also what is a "sysadmin"? I rarely see adverts for that any more, as most orgs have moved to DevOps / SRE model for well over 5 years. Also, it wasn't COVID that changed anything, it was lockdowns. It's important to get that straight.
Also, it seems like this has come from academia where they probably still do "sysadmin". I remember working for an academic entity in 2012 and trying to convince them to adopt a DevOps approach, to which the team leader asked "If we automate all this stuff, what will my team do?"
:facepalm:
(of course they would be able to do MORE stuff, like better security, automated deployments and look after more systems - there's always work - but perhaps not for the DBA any more, who have all but disappeared)
《A Victor Meldrew reference!》
When I recall Victor's ultimate demise and realize now I could pass for a pretty fair replica, its scary.
When you scrape away the (bull)shit(ters), sysadmins are the only ones that have any idea at all how any/most of the technology works and how it breaks and hopefully how to fix it when it does.
Whatever flavour de jour, model or whatever is in use to manage complex systems, out in the real world, I cannot say the litany of failures reported el Rego's columns or more mainstream media, fills me with much confidence.
The trouble with any system (or desktop for that matter) support is that you are only noticed when things go wrong. I know that this is utterly unoriginal but it is still true. Back in the day, they noticed if the punch cards were not sorted properly. Now they notice if the auto-scale doesn't scale. Personally, I was too late to the party to experience punch cards and my current gig is not ready to step into the cloudy utopia. Plus ca change!
I am not sure what the solution is though I am already counting down to retirement.
Someone said that as they moved towards the devops approach (during covid) , they changed the way they worked. For example if you wanted a new column to your table, you used to have to walkk roun, to, and put in a request to the DB2 admins.
Now developers have the authority to do it themselves, when they have got it right, then they raise a request for the db2 admins to progress it into the pre-production system. This had benefits that development/test could make as many mistakes as they wanted (whoops wrong data type, or that "new index didnt help") and Db2 admins had less work to do, and there was less time pressure( we want it now!).
One person in each development team was "trained" on the DB2 work and had the authority to make changes.
Of course after the first few times when development destroyed the database, they learned to be more careful.
This - spontaneous conversations, helping each other etc.- so much was lost from my workplace when we al flew the coop.
In my org. (YMMV) virtual has never allowed for the same fluidity of just jamming with a whiteboard and a few markers to work through a complex problem/bounce ideas around.
Had a good exmaple of that the past weeks. I've been dealing with an issue with a newly installed Exchange 2019 environment. Even had an MS case to get it to work properly, but in the end their help didn't resolve the matter either. I've been talking to colleagues to get to the root of the issue, but it was just one on one talking. Today we all came into the office for a combined session, where all the involved engineers were available to check stuff. After some brainstorming and checking stuff that should be correct, we identified two issues that were at the root of the problem, everything started working as expected pretty much as soon as we changed the needed config.
I really think we could've nailed this down much, much faster, had we not all been working from our own, private islands.
And you couldn't all get together using Teams / Zoom / Slack / whatever, and looking at a shared screen?
I'm not sure "brainstorming" works any better using a whiteboard compared to a Teams call.
I think with WFH set to continue, people are going to have to adapt. Even if it's only 10% of the workforce who choose to WFH most of the time, you'll still need to collaborate with those 10%, so a whiteboard session will just not cut it any more.
Also, it's easier to record a "virtual" meeting.
I remember once having to do a very complicated install, and the expert flew out from the US to show me how to do it. He then sat a few desks away in the same office in Cambridge and talked me through the install by using a screen share... which was BETTER than having to pitch up next to his desk, and look over his shoulder.
I've been talking to colleagues to get to the root of the issue, but it was just one on one talking.
That seems to have been the problem. Could a virtual meeting of all parties have solved the issue as a physical meeting did?
I don't know. It seems to me you have demonstrated meetings can be better than one-on-one but have not proven that meetings need to be physical rather than virtual.
I am not convinced 'work from home doesn't work' based on just a few specific cases.
You could have asked you coworkers before going the MS-way. OF COURSE, that is easy to say for Captain Hindsight :D.
"Others said remote work meant the pressure to do things because someone was waiting for them to happen was not felt when they worked from home"
Personally for me once its time to go home, unless its something actually critical to the business I was out the door (1 minute delay = 5-10 extra traffic getting to the M25).
Nothing has changed for me working mostly from home, I do tend to login earlier (But then I always headed to work 10 minutes early to avoid traffic and so I could get myself a cuppa).
The problem with that is that there is so much informal, non-ticketed work that your jobbing sysadmin does on a day to day basis. From the "Help I've rearranged the columns on my spreadsheet and don't know what to do" type idiocy all the way to the "Huh, that machine drops off the network every 3rd Friday but only for precisely 3 minutes... what the hell is causing that?" kind of oddness that you happen to notice while glancing at the logs. There is an absolute shit ton of work that it's easier to just poke your nose in and fix now than raise a ticket, put it in a queue, and get back to it eventually, usually after you've forgotten why it was important in the first place.
As the study (for all it's faults) points out, when working remotely all that previously "I'll sort it out right now" work that was never tracked now leads to support tickets being raised, and therefore all the extra time spent actually handling the fact that there's a support ticket raised on top of the doing of the job in hand.
Just looking at a ticket queue that has 300% more tickets in it than you're used to - especially once that queue gets big enough to no longer fit on a single screen - has morale implications and can lead to choice paralysis when faced with a couple of hundred fairly trivial tasks all of which need doing, but none of which are actually terribly important right now.
Where I was working it wasn't a case of "overwhelmness" but more a case of notenoughness and toldyousoness. Not enough dial-in bandwidth, not enough remote-enabled accounts, not enough VPN licences, not enough standalone licences for development tools, etc. and IT reminding the high-paid help that all this came up in the risk workshop at the last off-site shindig.
3 years I've been working remote, with two different companies, and it is obvious to me that this isn't about tech, but process.
My first company went into remote from an office environment, and it was no different - in fact, it was better.
We did the zoom thing, with loads of screen shares, and it was just like the office.
Then I move to a company that had been remote for ever.
They don't do loads of zoom, and everything is totally formal, ticketed, and issues are solved with loads of meetings with extra irrelevant people.
Culture is what you need to develop, and that is true for for whatever tools you use.
Next job is with a hybrid - we'll just see how they do things, but as its still in a secure formal environment, it will be interesting to compare notes.
Way back when, when I was a sysadmin, there were pretty much no times I'd need to be within 6 feet of another person, or even 60 feet. So I wouldn't have been worried about catching covid, and not allowing walk ins in my office and getting interrupted would have improved both my mental health and my productivity!
And it is the response to COVID which caused all the changes, like you say! So Covid did change it. Tough it is not "Covid's fault", as you write, since Covid is just a virus without will or intelligence. It is humans fault to allow that thing to jump from animals to us.
Without the response and our, relatively, high vaccination level we would have had it much worse. Check the current statistic worldwide. Tough you should click on "yesterday" and "2 days ago" since the current day statistic lags, by design, behind.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
