Re^2: Drowning in Modules - Suggest those for DB App

[gellyfish]

Plus, I would strongly suggest not using Windows for a business-critical server. Most Windows admins build a reboot into their standard administration protocols. To me, that's a red flag.

This may have been true five years ago (and don't I remember those early versions of NT 4 with horror ...) but I don't believe it is an accurate representation these days. The project I am currently working is implementing an entirely Windows based Billing system for a multinational company with thirteen operating companies in 9 different countries, this kind of thing is quite costly and failure is generally not considered to be particularly career enhancing, so understandably the people making the decisions are extremely cautious when choosing the software and the platform it is to run on. Some very intelligent people have bet their jobs on the reliability of the application. This system achieves in excess of 99.5% availability across eight or more windows servers.

Windows might not be suitable for every application but to suggest that is not suitable for a "business-critical server" is palpable nonsense.

/J\

[dragonchild]

How about this as a compromise: Maintaining a Windows server to given level of reliability takes more experience than maintaining a Linux server to a similar level.

[gellyfish]

I would put it slightly different: Maintaining windows takes more experience than is generally available to those often given the task of doing so, whereas those charged with administering a Linux (or other Unix-like OS) will generally have the experience of (or inclination to learn) the superficially more difficult skill.

In my experience Windows Server Admininstration is often foisted on some unsuspecting desktop PC support guy, who,lacking in experience on the server platform and when presented with some out of the ordinary problem, will fiddle around in likely looking bits of the Control Panel and Administrative Tools until the thing starts to work. And the people wonder why the server becomes unreliable.

/J\

[dga]

99.5%? Thats nearly 2 days of downtime a year. I think that most business critical systems would normally be defined as 5 9's (99.999%) which is just over 5 minutes of downtime per year. I have run 'not really critical systems' in the past where the expectation is 4 9's (99.99) which is less (<1 hour/year) (~3157 seconds) and at 3 9's (99.9%) (<9 hours/year), I would have been asked to 'seek opportunities elsewhere'.

My current server system isn't so critical but it has well past 4 9's and I only admin it very part time. Needless to say it doesn't run a Microsoft OS.

In fact it has had less than 2 hours of unplanned downtime in 5 years and thus I would consider it 4 9's.