Hey Admin, do you think this is your responsibility? [closed]

Question

I use linux all the time and I can manage several server related tasks. But I'm employed as an web developer and we have an admin to manage several virtualized nodes (i have access to most of them too). I'm wondering how much responsibility the admin should have to keep the application running or at least to communicate or see upcoming issues.

Example: We had to change the URLs of our application. He set an alias for the new domain name within the Apache vhost config, but he missed that there was an mod_rewrite rule containing the old domain. Which, of course broke our application.

How do you as an admin handle this problem? Do you try to learn how the application works and talk back to the developers so you can find the best solution, or do you usually push application specific issues to the developers?

Answer 1

As an admin, my responsibility is what is outlined in my job duties. If your employer doesn't have any set of rules assigned and simply wings it, then you kinda' get what you get.

Why? Because it entirely depends on your business. Admins could be responsible for a specific set of applications or servers. Or he could be responsible for servers and workstations. Or just keeping user workstations working. Or just keeping a database farm working, and someone else keeps backups working, etc.

If you are having problems with what should be whose responsibility, you need to talk to the person in charge and get things ironed out. There is no single set of "should be" responsibilities for IT that fits for every business.

Answer 2

To add - pretty often it is not the job of the admin to check for further changes. Admins get for updates etc. a SOP (Standard operation procedure) to follow. If you want further changes, then WRITE THEM IN THE CHANGE REQUEST.

Specifically, admins keep servers running and run standard tasks on them. Any ugprade etc. is not decided by the admin. Any change done for custom stuff nees to be exactly scripted for the admin to follow (not as technical script - I mean a word document tellnig him exactls what to run when). The admin in general is NOT responsible for the correctnes of that - because he OFTEN does not know wnough (keeping 300 servers running is work enough - I dont really have the time to track all changes and the reasons etc. for 100 busienss apps on the side, the app owners plan that).

So, in the example, you should ahve provided an exact outline what work needs to be done on what server and it is YOUR responsibility to make sure you cover all rewrite rules. The admin may not even know / remember that rewrite rules are being used.

Answer 3

In general, whatever the work description says.

In this specific case ("change base URL", needing done in multiple places), "it depends". If the change request contained just "change the 1.example.com domain in the vhost entry to 2.example.com", the change request is broken and whoever wrote it is "at fault" for the production outage.

If the change request contained "change the base URL from 1.example.com to 2.example.com, following the App17 design document" AND said design document actually contained all the places it needed changed, then the sysadmin should've followed the design document OR spoken to the devs saying "this is a bit unclear, do I need to..."

If you take case #2 above, but there's nothing anywhere, it's clearly the devs at fault, as such a dependency should be documented.

Answer 4

I think you're confusing a System Administrator for a "Release Analyst" (for want of a title), but often the responsibilities between these two roles overlap.

As someone rightfully pointed out, you should have raised a CR, created a basic implementation plan (including roll-back) and had the changes deployed in an orderly fashion. You'd know all the changes required to all hosts because you've already captured the changes when rolling this out to dev/test/staging environments, right?

Answer 5

As it applies in my workplace (which may be different from everyone else in the world):

The Developer is responsible for creating an outline of everything that needs to be changed. Files to install, Apache config that needs to be changed, database updates, whatever.

The Admin is responsible for execution of that plan and knowing if/when it's necessary to deviate on account of external factors.

So in the example given. The dev should know that the rewrite rule exists, and should document that it needs to be changed. The Admin should make the change and report back to the dev what was changed and if there were any deviations.

We also have "Lead" employees who keep track of 3rd party program updates and notify the Admin team when necessary. This allows a clean division between who wants stuff done (dev/users) and who actually gets it done (admin).

This may not apply to your situation, as other have pointed out, you need to consult your job descriptions.