Dedicated server choices

[mtbc]

I am finally getting back around to the migration effort to unmanaged dedicated servers. However, a feature I much appreciate is always-on KVM / VNC / whatever access to the console as I do not want to have to boot from a rescue image and try to work with an operating system that is effectively on different hardware thanks to QEMU. So far I have found that reliable console access comes with only the more expensive server offerings, doubly unfortunate given that I also try to keep a second server for redundancy. I do not intend to spend nearly a hundred pounds per month for plenty of CPU and disk that I do not need; I doubt that I could recoup much of that cost through Gridcoin or similar. So, I shall sleep on the question.

Comments

[goldibehr]

If you don't have access to console, what do you have to do if your server isn't bootable? Does the provider give you web portal to force-reload a default image, or do you have to call their tech support?

Given the low probability of it becoming unbootable, it doesn't seem worth paying a lot. Would a little downtime really be a big problem in the worst case scenerio?

[mtbc]

Part of my issue is that I don't want a hosting provider's default image, I want to run the vanilla distro installer myself because I know it well and am sure of what I end up with after it is done; it also remains constant across providers. Many providers do offer short-term console access by request but I am generally not paying for fast out-of-hours support so, for example, when at around 6pm on Friday I found one of my test servers to be offline there was nobody who could help me.

Generally providers offer through their web admin panel the opportunity to boot from a rescue image which can suffice if one can then mount the regular filesystems and detect from logs and config files and whatnot what the problem actually is, but for my issue on Friday the logs hadn't been touched and the config looked fine to me. It could have helped enormously to be able to see what was going on on the console during the failed reboot. So, without console access my downtime can end up being days instead of minutes.