Replacing servers in the age of virtualisation

[pcstru]

The time of the IT Manager are, I believe, drawing to an end. The number of times infrastructure and strategic decisions have to be made are very limited. How often do you change wireless or switches? As for the on-site stuff, as stated previously, on the decline. You just have to look at the pittance that is being offered in salaries for an IT Manager to show how much schools value IT technical skills.

IT salaries in Schools have always seemed poor on average compared to the private sector. I'm not sure people here will recognise themselves as having poor technical skills (or thank you for implying they do) so perhaps schools just find that they can employ the people they need for the money they offer?

 

I'd agree that the IT manager is on the way out but I disagree on the reasons why. Larger schools may want more from their managers - so you may find yourself (as I have) managing other functions (I manage Science and ADT technicians, our library and reprographics.). Smaller schools may find they can outsource network and core service support - there must be a market for preconfigured puppetted core services, a form of IAAS (on a local or remote iron) and such a thing would be easier to scale support for.

 

As for a symbiotic (my word) partnership between IT professionals and teachers, please! Call me a cynic, but after 31 years with time on both sides this is an illusion. Teachers and I do have to cite the new ones in particular here, stick with what they know best and suits them. So getting them, for example, to move their "stuff" to O365 will only work when you make their home folders read only, tell them they have 6 weeks and remove the drive after that. As for getting them to stop using removable media... Symbiosis or partnership suggests there are benefactors on both side. There are not. I see it more as a oligarchical relationship.

 

When teaching staff understand this relationship, only then can we move forward.

 

It seems surprising that you predict the death of the IT manager when you (seem to) see the replacement as lacking competence or having the necessary breadth of knowledge. Schools might be a little blind to the importance of IT - but that only lasts until it starts to go wrong. Something that doesn't work for whatever reason, will not become the norm.

[Dave_O]

I'm not sure people here will recognise themselves as having poor technical skills (or thank you for implying they do) so perhaps schools just find that they can employ the people they need for the money they offer?

 

My experience is clearly different to yours. With the 20 or so school's I have worked with over the past 10 years all of the SLT teams, teaching and admin staff have described themselves as having poor technical skills none have taken exception when described as such. Quite rightly, their skills lies in the understanding of educational context not in their understanding of the technical nuances.

 

As to "perhaps schools just find that they can employ the people they need for the money they offer?", a telling retort.

 

 

I'd agree that the IT manager is on the way out but I disagree on the reasons why. Larger schools may want more from their managers - so you may find yourself (as I have) managing other functions (I manage Science and ADT technicians, our library and reprographics.). Smaller schools may find they can outsource network and core service support - there must be a market for preconfigured puppetted core services, a form of IAAS (on a local or remote iron) and such a thing would be easier to scale support for.

 

I'm lost as to why my reasoning is at odds with your stated rationale. Like you I manage all areas of information delivery across a MAT.

 

 

It seems surprising that you predict the death of the IT manager when you (seem to) see the replacement as lacking competence or having the necessary breadth of knowledge. Schools might be a little blind to the importance of IT - but that only lasts until it starts to go wrong. Something that doesn't work for whatever reason, will not become the norm.

 

 

I am unclear how you arrived at this conclusion. In terms of the replacement "lacking competence", I do not believe anywhere in my posts have I suggested that the replacement would be in any way lacking. I believe I used the term "e-learning educational specialist", representing someone who has the skills to help bring together technology and education, this should be the replacement for the IT Manager. What salary level this should be at, I don't know. As eluded to previously, high level technical skills are needed less and less in schools as we move forward.

[pcstru]

My experience is clearly different to yours. With the 20 or so school's I have worked with over the past 10 years all of the SLT teams, teaching and admin staff have described themselves as having poor technical skills none have taken exception when described as such. Quite rightly, their skills lies in the understanding of educational context not in their understanding of the technical nuances.

Your original post clearly referred specifically to "IT Managers" in relation to salary, not SLT teams, teaching or general admin staff. I'm afraid I can only respond to what you actually write. If my responses don't correspond well with what was not written, that is not surprising (to me anyway).

As to "perhaps schools just find that they can employ the people they need for the money they offer?", a telling retort.

Again, I'm afraid I can only respond to what you actually wrote. The "telling retort" is just pointing out the fact that Schools do fill posts on the salaries they advertise, salaries you contend are low. Since they do fill posts, perhaps it is not a case of schools undervaluing skills, it just is the market value for those skills. I'm not sure what might be telling about that other than it tells exactly what it says.

 

I'm lost as to why my reasoning is at odds with your stated rationale. Like you I manage all areas of information delivery across a MAT.

 

I am unclear how you arrived at this conclusion. In terms of the replacement "lacking competence", I do not believe anywhere in my posts have I suggested that the replacement would be in any way lacking. I believe I used the term "e-learning educational specialist", representing someone who has the skills to help bring together technology and education, this should be the replacement for the IT Manager. What salary level this should be at, I don't know. As eluded to previously, high level technical skills are needed less and less in schools as we move forward.

I understood "e-learning specialist" to be someone light on the kind of skills a Network Manager would have (detailed technical) , instead focusing on pedagogy so likely to come from qualified Teaching staff (my conclusion that you think them not up to this comes from your statement that Teachers "stick with what they know best and suits them"! - which is hardly a glowing reference to their ability to innovate). Perhaps you need to elaborate a little what you mean by an "e-learning manager", since it is not a well understood or common role at the moment? As a school, we have tried here with a Teacher leading on "IT across the curriculum" and it has so far been a bit nebulous and ineffective - at least as far as innovative use of IT in the classroom goes. IMO there is no question that the curriculum requirements are not the domain of an "IT manager", but we have more innovation happening with the subject domain experts (so within departments) than being driven by our (version of) an "e-learning manager". As a general observation, it's not lack of innovation that is difficult when it comes to effective IT use in schools, it is poor implementation (change/project management). Schools seem to be full of people that believe that if they say "we will do X" it will just magically happen. Even good ideas will fail to deliver value when poorly managed.

[dhicks]

I run rolling replacement for servers on a 5 year plan, not because the servers themselves are unfit at that point but because financial constraints mean that spreading the cost that way is nicely controlled.

 

A good point, and one which has me thinking that we could start to replace our aging hardware with cloud-based services, which are then paid for with a monthly fee rather than with a large fee every 3 / 5 / 7 years. Our servers are about the same age as localzuk's, it's time to start planning what our next upgrade is going to be. We are in central London, we have good access to high-speed internet, so running some / most of our server infrastructure on cloud system of some sort is a realistic option. Exactly what type of "cloud" is, of course, a question that will take a bit of thinking about.

[Dave_O]

1. I think we are talking at cross purposes here. I am pointing out that there is a diminishing need for high level technical skills. How you can draw the conclusion that this equates to me saying they have "poor technical skills" I have no idea. Please enlighten me.

 

2. This is what I hate most about the attitude of a "manager". I believe it is fundamentally wrong to pay as little as possible for technical skills. Just because the "market" ,whatever fickle beast that may be, thinks one should pay a pittance. I would rather value the individual. Do I value managerial skills more than technical skills? I do not.

 

3. You make many assumptions some of which are wrong a) E-learning specialist would not necessarily be a teacher b) I never used the term "e-learning manager" I use the term specialist (see 2 for my views on this matter) c) As to my comments about teaching staff, 20 years teaching and interacting with many staff across a variety of subjects and establishments has lead me to this. Innovation in IT from teaching staff is a very rare beast indeed.

 

4. "As a general observation, it's not lack of innovation that is difficult when it comes to effective IT use in schools, it is poor implementation (change/project management)." This is one area I mainly agree, although I would emphasise the importance of effective and appropriate training (one area perhaps for the e-learning specialist)

[jmak]

I've just bought next business day on site parts and labour warranty extension for our host server - for years 6 and 7! It wasn't expensive and it's direct from Dell. To me that indicates that they think a warranty claim is pretty unlikely. As all the servers are virtualized, it would be much quicker than in the past to change host.

[pcstru]

I think we have drifted off topic too much so have moved the discussion on IT Managers and stuff to : http://www.edugeek.net/forums/general-chat/168966-future-managers-schools-replacing-servers.html#post1446083]. Edited April 18, 2016 by pcstru

[Mr.Ben]

Back to the original post....

 

I think it's a difficult decision with a 2 node cluster, especially as you would be expecting expansion in the next few months. Assuming that the MAT expansion puts the workload beyond the capability of a single node:

 

Replace a server this year, and then one the following year - and expand the cluster to be 3 nodes (Treat the 3rd node as a fail-over node). Put both new servers on 5 year warranties (in 5 years time we may well have all moved to the cloud).

 

That way (assuming you have no failures in the meantime), you'll end up with a 'spare' server that can be raided for spare parts to keep the out of warranty node up and running in future years.

 

I don't think it would unreasonable to argue that you should only plan for one of the nodes to be down at a time. It also give you flexibility with updates etc during working hours.

 

As for the storage - Absolutely replace once it is out of warranty (assuming you only have one :-). Move the 'Old' SAN out, replace the disks and use it to store backups.

 

There are of course risks to this strategy - planned obsolescence on behalf of the server vendor being one (and it is the most frustrating thing!)

Edited April 18, 2016 by Mr.Ben

[pcstru]

A good point, and one which has me thinking that we could start to replace our aging hardware with cloud-based services, which are then paid for with a monthly fee rather than with a large fee every 3 / 5 / 7 years. Our servers are about the same age as localzuk's, it's time to start planning what our next upgrade is going to be. We are in central London, we have good access to high-speed internet, so running some / most of our server infrastructure on cloud system of some sort is a realistic option. Exactly what type of "cloud" is, of course, a question that will take a bit of thinking about.

I have to admit, having resisted 'cloud' services up to now, I expect now when I run the numbers moving guest servers onto Azure will start to become a sensible option rather than replacing iron, so in the 3-5 year period, this is likely what we will be doing. I also expect M$ licensing will be providing a hefty push in that direction too.

[CyberNerd]

I have to admit, having resisted 'cloud' services up to now, I expect now when I run the numbers moving guest servers onto Azure will start to become a sensible option rather than replacing iron, so in the 3-5 year period, this is likely what we will be doing. I also expect M$ licensing will be providing a hefty push in that direction too.

 

I think we are pretty much here too. At the moment the costs for 24/7 virtual machines machines are still quite high. There is a bit of a price war between AWS/Azure/GCE and Google at least have a commitment to a 'Moore's law' pricing philosophy so the prices are guaranteed to come down. I think a hybrid solution where we could boot up servers on demand is the way forward. This would work well for SIMS (if capita could support it) and for application servers. An offsite webserver is a no-brainer as the resources needed are tiny. I'll probably move at least one DC offsite soon and plan to have no servers left (aside DHCP and the like) in the next 5-10 yrs.

[mthomas08]

Must of missed this thread so looking through it now:

 

So, our servers here are 3 1/2 years old, and I'll be looking to replace them next year, possibly.

My question is - are people still sticking to their old rolling replacement program for their servers now that it isn't really necessary?

Note also, I'm not talking about storage - that's an entirely different matter.

 

Maybe just me but at that old I wouldn't be thinking of replacing them. Then again because I end up working in systems that start in a mess and spend years trying to get them to the standard rolling age of 5 Years, I've rarely had the chance to actually do it.

 

Personally I think with redundancy and reliability of equipment the standard doesn't apply any more. We have a Virtual Server that was expensive and it's now 3 Years old, it's had no issues so far and seems to be running perfectly still as if it was 1 day old. If parts started to fail often, that's where I would start thinking okay enough is enough. It has enough redundancy though that if parts do fail we can replace with little to no down time. I wouldn't consider this for replacement till it was at least 5 Years old but I would plan for it as a contingency replacement not something I WILL replace until it becomes a concern.

 

Our Hyper V is currently on 65-70% usage/capacity used and is working very well with zero issues. I think on it's 5th year I will think about it but so far, touch wood - no need to even consider it.

 

Storage though.. that really is a whollllllleeee different ball game.

 

That's how I feel towards PCs too. I need to be replacing what most likely will fail or the lack of performance out weighs costs. No point replacing something that could very well be used/useful and is working - especially with budgets being cut for Schools.

 

I also have a meeting today with Governors on proving what I do for cost control (this is important, anyone involved in heavy spending is having one to ones with the Governors).

 

I have to admit, having resisted 'cloud' services up to now, I expect now when I run the numbers moving guest servers onto Azure will start to become a sensible option rather than replacing iron, so in the 3-5 year period, this is likely what we will be doing. I also expect M$ licensing will be providing a hefty push in that direction too.

 

Likewise I am considering Cloud but as mentioned in other posts. Do I like the idea of our Film/Art Students and even a portion of our IT ones saving large files and downloading them to a "Cloud" based solution. BUT because I believe in being thorough even in systems I question - I have asked what costs/work would be involved in a possible new Broadband provider in sticking our internet connection on something fast enough to keep up.

 

Currently we are on 100MB, capacity hitting up to 75% an average due to the amount of BYOD/Internal devices we have. Now consider our internal devices no longer going to servers for data/email internally instead going through our internet. We've already upgraded our Internet from 50MB to 100MB due to capping it consistently (which is why we are still hitting 75% right after we upgraded).

 

Still I like to plan for a New Data server and Cloud based options. Backups, restoring emergencies and restoring Student work from 3 months ago will be on my considerations. Including our reliable the Internet is. I like the idea of saving us a lot of money though

Edited April 19, 2016 by mthomas08

[PotNoodleTech]

The days of needing to replace a server spot on every three years are over. I am not sure those days ever actually made it to most schools though.

 

Nowadays - you can easily run a server for 5 years, often more. Get a 5 year warranty thrown in on the server, it's much easier to negotiate when you're buying and playing a few manufacturers off against each other.

 

I would suggest that in your long term plans you plan to have the required budget to replace them every 5 years, if you end up leaving them till year 6 or even 7 then that means you'll have extra money in year 5 or year 6 to spend as and where needed.

[CyberNerd]

Currently we are on 100MB, capacity hitting up to 75% an average due to the amount of BYOD/Internal devices we have.

 

75% average usually means the line is saturated at peak. Saturated connection mean data loss. I think you should review your capacity planning.

[mthomas08]

75% average usually means the line is saturated at peak. Saturated connection mean data loss. I think you should review your capacity planning.

 

Already been there and done, even our current LEA Broadband Provider have been involved.

 

Unfortunately due to mobile devices and more "Online" reliance there is limitations to what physically can be done. Although we are reviewing our current broadband (as mentioned) to a possible 200MB connection and the big IF we go Cloud based. Things like this always on my radar.

 

For our current Broadband provider the costs are heavy for a 200MB connection, to a new one who are going to provide like for like it will be a 35% cost reduction.

Edited April 19, 2016 by mthomas08

[CyberNerd]

@mthomas08 I found that the main problem with BYOD was the windows updates. A caching windows update server might be a good thing to investigate, and restricting bandwidth to MS et al.

[Fazza]

Are the servers running fast enough for you? Are you running out of RAM and/or storage? If not, then stay as you are! If they are out of warranty then discuss the matter with your line manager and explain that the servers are running fine but if any hardware fails then they'll have to stump up some cash to replace the faulty hardware/server which is cheaper than buying new servers for the sake of it.

[mthomas08]

@mthomas08 I found that the main problem with BYOD was the windows updates. A caching windows update server might be a good thing to investigate, and restricting bandwidth to MS et al.

 

Not viable for us because we don't have control over our BYOD Windows Devices, but because as I said I like thorough. It's not Windows Updates nor is it Google Store for the Android Apps.

We use Android, Windows primarily for the BYOD but Students still have endless phones, iPads etc. We do however prevent iTunes through it.

 

Still it doesn't matter because we know the cause and the Scheme is being scrapped anyway, plug pulled and being thrown out the window.

Edited April 26, 2016 by mthomas08