Ah The Cloud
What a perfect excuse not to get any work done because the Rejects of Redmond buggered something again.
Microsoft 365 started the week off with an early totter in the Oceania region followed by a full-blown European Outlook client TITSUP*. The European failure appears to have blanketed much of the continent as Outlook users (connecting to the company's Exchange Online service) found themselves unable to connect. The problem …
Unfortunately I was not affected.
How I know?
Well, a co-worker dispatched no less than 20 (perhaps even 30) meeting invitations for our daily standup meeting. ...and she did that while I was busy clicking 'No, bloody fcking NO!' on her previous batch of invites. Watching the new avalanche of invites was surreal.
A mutual colleague tried giving her some pointers the first time this happened, and her reply was something along the line of setting up rules in outlook to filter incoming e-mail. Which he did. So he no longer reads anything dispatched by the mistress of mail missive avalanches. He also decided to move to another division of the company, making sure that his and her paths would never cross again.
I'm involved in a couple of projects where we're moving customers from Exchange on prem to Office 365/Exchange Online.
Having seen the state of these on-prems Exchange setups, I think the skills required to manage on-prem email (Exchange or something else) are rapidly diminishing. For many people cloud email (Exchange Online, gmail, etc) is the sensible solution. I used to be anti-cloud, but for many people, there just isn't any point in trying to run your own mail setup. It's a distraction you just don't need.
Is cloud perfect? No. But I'm a pragmatist and even the occasional cloud issue is minor compared to when a customer hasn't done any maintenance to the email server for years and finds their server then goes belly up for days as we come in to try and rescue it.
Precisely. For £500 per year we can get an email system and document system with 1TB storage per user for a company of 10.
If anyone can beat that price for an on-premise system I'd be interested to know. We don't need 100% uptime, and to be honest we'd probably settle for 99%. Bear in mind though, that your fee for fixing it in the 7 hours per month that it can be down needs to be covered by your quote. We're not paying for you to fix it!
And these days, is e-mail your only communication platform?
My employer uses both Teams as well as Slack. And Yammer and Confluence and lots of other platforms where some form of communication can take place independently of Exchange.
I'm no business manager, but 99% availability for e-mail sounds plenty good to me.
"Exchange Server crashing quietly in the corner"
There were no major issues with the Exchange environment I have managed for the last 10 years that stick in my mind. I am sure there must have been some minor problems but I can't recall any crashes or mail outages. It has moved from 2010 -> 2013 -> 2016. There were some high resource usage issues with search and a few connector niggles with 2013 when it was first released, but since SP1/CU4 it has been solid. Redundant client access and DAGs meant upgrades with no downtime. It just worked.
We only just moved to 364 due to the absurd minimum resource requirements specified for Exchange 2019. Just in time for mad Monday.
Tips for happy Exchange Server-ing.
* Have a healthy AD and DNS infrastructure.
* Have at least two Exchange servers with a DAG and load balanced client access.
* Give them lots of disk space. (No really, give them double what you thought they needed to start with.)
* Plenty of RAM as well.
* Apply quarterly updates a couple of weeks after release.
99% of sick Exchange servers I have to deal with fail the above. Usually the first one. Quite often they are resource starved as admins assign resources based on experience with previous versions of Exchange. 2013 onwards are a completely separate beast that log absolutely everything and cache as much as possible to RAM.
If you cant have more than one Exchange server then you need downtime every three months to apply the CUs. Or you just don't bother (usually the case).
General rule of thumb, if you can't have more than one server go to 365. If you have to stay on prem, get another server.
Come on your must have accidently caused some issue, me as Domino admin found one of our sites on a seperate notes network, corrected this and minutes later rather found lots of leavers staff re-added. After wondering what went on discovered a backup of names.nsf was taken and left in the data directory, as the deletion stubs for the entries were long gone they were re-populated.
Ah the joys of dodgey setups.