What is up with the attitude from sysadmins? [closed]

Question

Every time I ask a simple request like spinning up a new server 2008 VMWare instance, these guys act like I'm asking them to build the great wall of china overnight. I've had to do this myself many times, it takes under 30 minutes, and maybe 30 seconds of user interaction.

Why do infrastructure/sysadmin types have this "say no and delay everything as long as possible" attitude. It drives me nuts and doesn't let me perform my own job tasks (developer)

Answer 1

I'm a developer by trade, and there have been times when I've felt similarly. In my case, I've found out that there are usually wrong assumptions on both sides of the discussion.

To illustrate, here are some things that SysAdmins at my work consider when handling a "simple" request like spinning up a new virtual server.

• Licensing for the operating system - is this a reallocated license from somewhere else, covered by an MSDN subscription, or a new license to be tracked?

• Security and reliability patches need to be pushed out, and the machine needs to be added into the inventory of servers to be monitored and maintained.

• Where is the hard drive for the virtual machine going to live? Is there any information on that drive that needs to be backed up? If so, can we move that information out of the drive onto the SAN for replication?

• Is this going to be a development box, for a testing environment, or for production?

• For a development box, what software is going to be installed? Does the machine need to have internet connectivity? If so, what about Anti-Virus and other malware? Who's tracking the software licensing of the development tools?

• For a testing box, are the licenses covered by MSDN (in our context, it is for system testing by developers) or do we need more licenses for all the software installed because it's not used by developers (in our context, business testing).

• For a production box, what is the SLA for the machine? Does it need to be replicated to offsite? Do we have room on the backup tapes for this machines hard drive? Does the DR site need to be a hot-swap, or can some down time be tollerated? If down time is Ok, how much? 30 minutes versus a week - different solutions required.

• Who needs to have access to the machine? Will they be using the VMWare/HyperV console access or Remote Desktop (RDP)?

• How long will the machine be needed? Who is responsible for the machine? In a years time when no-one remembers what the machine was for, can we just turn it off?

Keep in mind, I'm not a systems adminstrator - I'm a developer. These are just the issues I'm aware of - the issues I've learnt about. I'd be very suprised if the list was comprehensive or complete.

How did I learn? I worked hard on developing a good relationship with the SysAdmins, and by going out of my way to help them help me.

In every case I've experienced any attitude I received from a SysAdmin resulted from a lack of clear communication. Being clear what I wanted, and why (keeping to information relevant to the SysAdmin), being willing to listen, and being willing to learn about their concerns and working within their constraints were all important.

Answer 2

Some outstanding answers here already, but to add:

Check out the beginning of Tom Limoncelli's "Time Management for System Administrators". Read the Preface and Forward (maybe 10 pages altogether), and maybe CH1; you should be able to read that much on Google Books or Amazon...

I've never met a dev who's work was interrupt-driven, but as a SysAd (who can and does code) I've never had a job that wasn't interrupt-driven but also required daily maintenance, break/fix, and project work. To really understand why "it takes so long for something so simple/fast", imagine trying to read a book with somebody poking you and asking a question at random intervals (say between 3 and 30 minutes apart.)

It may only take 30 minutes to fulfill your request, but does it need to happen right now? When did you know about your requirement? Is it really urgent enough that I should drop what I am doing? If you knew about it earlier, why didn't they ask sooner and provide a (reasonable) deadline?

Answer 3

Because yours is a single request in a sea of requests that they take from a lot of people and not just from you. Every time a single request arrives priorities and queues change. Yours probably is not an interrupt driven work, but you have to accept that when one's work is interrupted, stopped, switched to another task due to higher priority, sometimes the attitude may not be as polite as you might expect.

Users rightfully expect their requests fulfilled immediately as they enter the queue, and do not care about the queues length and priorities. The hard fact of life is that they exist and they are managed by system administrators. So yes when when someone does a work of low visibility (when everything is OK isn't a sysadmin redundant?) and high impact (when something is broken will it be back online in ten minutes because we are in a hurry people!) some people develop attitude.

Delay is the result of a long queue, coupled with your sysadmin's actual "speed". If you want them to improve on that, or locate causes for the delay, have them print data from their trouble ticketing system. That way you will help both them and you.

Answer 4

I usually ask - Why?

While this may be a virtual server, it is akin to asking a sysadmin to install a new server into the rack. The rack-space must be found, network ports must be connected, power needs to be pulled, software needs to be installed, licenses need to be configured, switches need to be managed, firewalls need to be configured, yada yada...

Installation is easy, management is more difficult. Spinning up a new VM increases the number of machines that need to be administered and controlled, that is all. Even if it is a virtual one, the techniques used to administer a VM is quite similar to the techniques for a real one.

However, if the request is a legitimate one, there is no reason why the SysAdmin should not help out.

Answer 5

If it's easy to do yourself, then do it. As long as you support it. But for a sysadmin, it's not quite as simple.

• Have to have resources on the vmware host (memory/cpu/hdd space) for the
  new server.

  Licensing for the software

  Configuring it (Firewall, required software)

  Add it to the backup system

  Updates Monitoring

  Documenting everything about it.

Then support it. (AKA, restart it everytime you hit shutdown by mistake :) )

If I need a server up to test something for a few days, I use GoGrid. Then I don't have to deal with anyone else and if I nuke it, it's my fault and I have to fix it.

Answer 6

Because your companies workflows suck? Sorry to say it so badly, but...

• There arae self service portals available. Hyper-V / SCCM can handle this easily automatically. No need to even get a sysadmin involved for a dev / tesst system there. If you use visul studio, Lab Management goes even further with virtualization management. They also handle ressource budggeting (so you dont use up all servers without remembering) and can automaticaly dear down forgotten machines (after like 30 days) unless manually set to let them survive. VERY nifty.

• There are dev servers in many companies where develoeprs have a pool of servers they handle VM's on themselves.

Basically, especially for small test / dev type servers, IT should not have to get involved manually.