5 ms·
Vampires (Programmers) versus Werewolves (Sysadmins)
- nkassis 16y agoSo considering I do both am I a Werepire ? But seriously, on small teams, you can't avoid that you will have a few people doing both for a while. We have a dedicated a sysadmin but we work with him a lot considering he has to deploy our code and devs are terrible at helping out sysadmins. I get in these arguments all the time with other devs about making sure it's easy for the sysadmin to configure and deploy the app. When they write code that depends on random libraries, they need to ensure that will be deployable on production servers running Linux and not OSX etc.. I feel having experience with both is important for sys admins and devs to understand each other.
- mgrouchy 16y agoChances are when you work on a very small team you probably don't have a sysadmin at all. At the startup I work at, I do both and the more I work without a sysadmin the more I appreciate the work they do! When we do have a dedicated sysadmin, at the very least he will be very appreciated by me, because I know he will be improving all the stuff I have done out of sysadmin ignorance.
- sophacles 16y agoI think it is extremely important to have at least one developer who has been a sysadmin in the past. One of my closest colleagues started his career as a sysadmin. At first I was a bit distrustful because he didn't have a lot of "real experience" or "education". But as I got to know him, I realized his sysadmin "spidey sense" was a godsend. He has a knack for knowing when things that should just work will in fact fall over. He also has a lot of deep OS and platform knowledge, so frequently when we would want to throw piles of code at something, he would suggest and OS tweak or similar to get the same end result. So yeah, sysadmins on all development teams from the get-go are now a must in my book.
- mgrouchy 16y agoYeah, I was never a sysadmin, but for 8 months in my university days, I worked as a systems programmer for the CS Department at my university. I worked on stuff that automated system deployments, backups, imaging, etc. Its actually in the process of being open sourced now at http://www.labnet.ca/ http://www.labnet.ca/ (Its actually a pretty incredible system I recommend anyone who is interested in this type of stuff to check it out). Because of this I worked with some awesome sysadmins and unix/linux gurus so even though they were a bit rusty, I had enough core systems admin skills to have a good intuition about how to solve most of the problems I have normally. When I run into roadblocks I know some awesome sysadmins that I can hit up to get me over the hump.
- dedward 16y agoWhy not work with him to build an automated deployment system? Get the group to think a bit bigger - consider deployment (and possible rollback) a function that is an integral part of the overall system. Get the sysadmin working on system stuff and automated provisioning tools so environments can be replicated and performance tools, monitoring, etc - and have the developers understand that they are building something that has to be deployed by this tool. The decision as to who gets to push the "go" button is one the group can make - maybe it needs signoff by the sysadmin, maybe it can be overridden with a warning if necessary - etc. These days it's cheap, especially in linux land, to manage and deploy proper development, testing, and production servers with automated tools -and time is better invested generally in working on those tools than manually provisioning things. It's a bit more work at first, but pays off hugely farly quickly.
- hcayless 16y agoThe big takeaway here is that if it's really "versus", i.e. the programmers and sysadmins are having a power struggle, you're screwed. If one side or the other "wins" you're even more screwed. If your programmers and syadmins are working at cross-purposes, then you have a serious problem that needs to be corrected immediately.
- arethuza 16y agoI often think that most of the benefit you get for making changes to production systems is from getting someone else involved when you make that change - whether it's an incredibly formal process (which can be justified for some systems) or "want to sanity check me making this change" it all comes down to getting other people involved so they will say things like "You do have a backout plan in case this screws things up, don't you?".
- boredguy8 16y agoIf the only thing your sysadmin is doing is 'looking over the shoulder' at code they may not (probably don't?) even understand, you need better sysadmins. You should be sanity checking with peers, whether that be other devs for a code change or other sysadmins for an environment change.
- shin_lao 16y agotl;dr A lot of noise to say "programmers and system administrators have a different job and they need to learn their differences to work together".
- jasonjei 16y agoThe problem with this article implies a bureaucratic division of labor when developers are often the sysadmins. Any upcoming startup can't afford to be resistant to change--and things will break, but nobody will know more about the software than the developer. This kind of thinking reminds me of the way the enterprise software world is run, and they're not exactly renowned for quality, either. I've found that great scaling is the result of sending sysadmins to work along side developers, and having developers work with production/testing to learn about each's side struggle. It's not R&D vs Operations, they're on the same team! If you're building a house, you might have architects and builders, and while the architect draws the blueprints, the actual construction people are just as important since they have an equally good understanding of the other side, such as applying the materials in good form.
- mcknz 16y agoAnother PSA for Stack Overflow.
- boredguy8 16y agoFirst, I highly recommend "Deployment is just a part of dev/ops cooperation" http://www.kitchensoap.com/2009/12/12/devops-cooperation-doesnt-just-happen-with-deployment/ http://www.kitchensoap.com/2009/12/12/devops-cooperation-doe... Second, this isn't a question unique to the dev/ops relationship. What about the dev v. project manager? If you have a good team of developers, why have a PM? Having worked as a dev, a system engineer, and newly in the PM department, everything I've done since being a 'code monkey' is motivated by a desire to make the dev's jobs easier. Maybe that's because of my roots, but I think it makes the most sense. Where I work, we have a useful distinction between Sysadmins and System Engineers. Broadly speaking, admins are the i-dotters and t-crossers and do things with well-defined processes while engineers are responsible for everything else. As an engineer, I wanted to make as much information about the production environment available to the developers as possible. I also worked to make sure the dev environment would mirror production as closely as possible given that we only had one dev sandbox (requests for a staging environment were declined). These two things, combined with being a good teacher, meant I rarely had requests for CLI access to any of our environments.
- brazzy 16y agoWhat about the dev v. project manager? Bah! Devs vs. DB admins, that's where the blood stains the rivers red, the crows feast like kings, and legends are born!
- jeebusroxors 16y agoAnd here I thought it was sysadmin v DB Admin. I guess no one likes the poor DBA...
- scrame 16y agoRight, nobody likes DBAs.
- ww520 16y agoIt's simple. Give developers access to production servers but not root privilege. Deployment and configuration change should go through change request ticket, whose main purpose is communication to the group and maintaining a history of changes to the production site. Change tickets can be as simple as sending an email to a group list or an elaborate system of reviews and approvals using ticket tracking app. Implementation of the change ticket can be done by developer, qa, sysadmins, DBA, or ops, depending on the needs and required privileges. Routine stuffs should be automated. It usually takes a few times of site down and screwup before some form of process to be put in place.
- CountHackulus 16y agoI find it interesting that he seems to disagree with his co-founder on what seems like a key issue moving forward.
- dedward 16y agoDevelopers should never have system level access. They should never be able to meddle with the core infrastructure their code runs on. This has nothing, however, to do with who should be able to publish new code to the website (we're talking web apps here, right?) - that is something that should be handled by documented process, and automated tools to ensure accountability. Hypothetically, as the senior sysadmin, I'd like to sign off on every change that's going public, because I'll have to deal with the fallout if something is wacky. Realistically, I'd also like our senior developer, or failing that, even a more junior developer to be able to push code to production in an emergency if the right people are there to approve it (perhaps not hte proper technical staff, but they can get the okay from the owner to go-ahead). Accountability, automation, and process are key here. The deployment system should be accessible to some people who could potentially get fired for using it wrong, and it should be very clear to them what the right thing to do is. The deployment system could even warn them of this and include official signoff, etc.