Prometeus 128MB KVM, 2 year anniversary
- Thread starter willie
- Start date
Haha Agreed.
256 MB KVM. Those restarts are my own restarts I did (upgrades plus moving things around here and there). As you can see the max before the biggest restart was 495 days.
Even though Italy is almost on the other side of the world for me, I love Prometeus!
My use (the entire time) has been my IRC Bouncer. Therefore can't really say much about the performance (hopefully you can chime in on this!), but everything seems great for my use!
All good things come to an end. I got an email from Prometeus this morning saying there had been a big power outage at their Milano campus earlier in the night, and they had staff on site putting stuff back together. I ssh'd into my vps and sure enough it had been rebooted, so its uptime was just a few hours (it was still working fine of course). Oh well.
I took the opportunity to dist-upgrade to debian 7 (it was running 6 before) and rebooted again. 2 years and 8 months of uptime, certainly the longest I've ever had any vps keep running nonstop. And it's a kvm: I think there are some situations where an openvz node can reboot or a container migrate without the container itself seeing the downtime, AFAIK the uptime of a kvm means it's been running without interruption. Wow!
I took the opportunity to dist-upgrade to debian 7 (it was running 6 before) and rebooted again. 2 years and 8 months of uptime, certainly the longest I've ever had any vps keep running nonstop. And it's a kvm: I think there are some situations where an openvz node can reboot or a container migrate without the container itself seeing the downtime, AFAIK the uptime of a kvm means it's been running without interruption. Wow!
DomainBop
Dormant VPSB Pathogen
It's not just OpenVZ. KVM and Xen guests can also be suspended before a node reboot so they will still show the same uptime after a node has been rebooted.
Here is an example where a KVM guest was suspended prior to the node being rebooted and after the reboot still showed 11 days uptime.
...and here is a Xen example. The guest was suspended prior to the node being rebooted and after the reboot the guest still showed 106 days uptime.
TL;DR the uptime command on VPS guests doesn't always reflect reality.
I got that email too. There are still a few nodes and services down (the MYSQL Offload server is still unreachable). IWStack seems to be the service they're having the most problems bringing back up.