So I was informed by support ticket 9 days ago that my Linode would go down for a *maximum* of 45 minutes today for scheduled maintenance.

So far it has been 2 hours, and ticket updates saying

- "maintenance has now completed"
- "We are investigating some instability with this host at this time"
- "This issue should be resolved at this time"

And my Linode is still down. Two hours! I use my Linode for websites to get new business. On a bank holiday as it is here in the UK (high footfall for new enquiries) this translates as lost sales for me. So unimpressed.

The worst thing is that I've updated my ticket twice now since they responded, which was now 35 mins ago.

I just want to know what the problem is, why they say it's resolved yet my linode isn't booting, and when I can expect the problem to be *properly* resolved!

Am I alone in finding this unacceptable?

Having a bad system architect that includes a single point of failure that directly impacts your business. So unimpressed.

Sure, Linode should explain whats taking so long, but if you're losing money, it's your system design that failed, so it's YOUR PROBLEM for having a single point of failure.

It's called R E D U N D A N C Y, and if a few hours of downtime actually costs you significant money, maybe you should look into it.

_________________
Either provide enough details for people to help, or sit back and listen to the crickets chirp.
Security thru obscurity is a myth - and really really annoying.

It's unfortunate, but shit happens. Anybody relying on a linode (or any server for that matter) for financial-related things (like sales) should have a contingency plan in effect for these sort of situations. For example, a second linode as a hot failover, or even a cold failover that you can spin up for the duration of the migration to limit downtime.

Last time we had a worry about such a thing (the hurricane that was near the Linode DC), I cloned our linode to Fremont for a few days, just in case. We didn't have to use it, but if the New Jersey datacenter got wiped off the face of the map, we had another linode to spin up and could have been back up and running in minutes.

I migrated host to an already-maintenance-performed host ahead of time when I knew I was going to be around at a time when traffic was lowest and it would impact least amount of people. Linode support were fine with this and migrated me at the time I wanted to be done.

Edit: Total downtime for the migration that I asked for was ~20 minutes with a Linode 1024.

I hear what you all are saying about single point of failure.

The issue I think is one of expectations. A business you pay money to tells you that your server will be down for a *maximum* of 45 minutes, and it's down for just over 2 hours in the end. And you aren't updated.

If I had been given the right expectation, then perhaps I would have put in place some failover.

In fact I had a spare Linode sitting doing nothing at the time, so I could have done something about it. I just didn't think it was worth it, on reflection to spend hours setting up a failover server when 45 mins I could stomach.

Am I wrong to trust a business when they tell me something will take a certain amount of time, and it takes over double?

For my clients, if I tell them about an outage, and it becomes apparent it will last far longer than expected, I will keep them updated. It's just good business.

You still don't get it - life is full of random problems.

In this case, Linode estimated one amount of downtime, but reality was something different.

Next time it could be an uplink failure, hardware failure, DC failure, power failure, asteroid hit, Godzilla attack - who knows what next time will be.

You can't predict when things will go wrong, you can only plan what to do when they happen.

You don't seem to have a plan for when things go fubar, except of course to complain that it's costing you money.

Make a plan, then it doesn't matter what happens, you're prepared and you're NOT losing money (which is what the game called "business" is all about).

If downtime is an expense you can't afford, plan for it - and I'd recommend having your backup server not just in a different DC but with a different hosting service as well (separation is your friend).

_________________
Either provide enough details for people to help, or sit back and listen to the crickets chirp.
Security thru obscurity is a myth - and really really annoying.

Weird, I've been with Linode for over 2 years, any time I contact support, even on the week-end, updates and getting things done are what they are best at.

Hey vonskippy, thanks for the life lesson! Before I read that pithy recitation I was clueless about how life can throw you lemons, now I'm thoroughly enlightened and it's thanks to you!

In seriousness - and I did explain this pretty clearly in my previous post - my issue with the Linode was somewhat *unrelated* to the fact there were technical problems.

I don't mind problems. "Life is full of random problems", as a wise teacher once taught me.

It's how you set expectations as a business. And then if something falls outside of those expectations, it's how you deal with communications. e.g.

1/ Linode could have aligned my expectations by simply appending the words "unless we hit problems" to the phrase "this will take a maximum of 45 minutes".

2/ When the problems happened, rather than keeping quiet, I would have found it far more agreeable had support contacted me with some basic info about the cause of the problem.

3/ More importantly, they could have provided me with some reliable information; rather than telling me that the problem was solved, perhaps they could have verified that my machine had indeed come back up. Instead, I kept trying to boot the machine for the next c. 45 mins thinking maybe I had done something wrong.

4/ So yes - after telling me the problem was solved, it took another 45 mins if I recall for them to come back and tell me that the problem had not been solved and there were still instabilities.

Look, I'm not really asking for an answer now... but I must admit I found your response a little antagonistic hence I felt the need to respond in this way.