Re:Updated Linode Plans / New Larger Linodes
by remyM on Thursday May 31, 2018 @07:55PM
(#17201366)
damn, still no Australian DC.
I'm stuck on Vultr still, since they're cheap and they have an Australian datacenter.
AU routing to JP is really bad (I get faster connections and better ping to TX-US servers on a lot of providers), and SG is like 120-160 ping which still kinda sucks.
Re:Updated Linode Plans / New Larger Linodes
by fiveFeet on Thursday May 31, 2018 @02:07PM
(#17198587)
Are the cloud hosting services profitable at these very low prices or are they reducing the prices just to gain market share? If it is the latter, sooner or later they have to increase prices, no?
Re:Updated Linode Plans / New Larger Linodes
by neom on Thursday May 31, 2018 @02:43PM
(#17198913)
I would imagine most clouds will have to be pushing towards primitives (store, compute, network) being lost leaders. margins are on value-add services (software), deploy, monitor, scale etc.
Re:Updated Linode Plans / New Larger Linodes
by organsnyder on Thursday May 31, 2018 @02:19PM
(#17198705)
Linode has been around long enough (15 years) that I can't imagine they're operating at a loss. Their prices also haven't been nearly as dirt-cheap as other providers.
Re:Updated Linode Plans / New Larger Linodes
by andypants on Thursday May 31, 2018 @01:22PM
(#17198192)
OVH is not good at reliability. Entire data centers go down and there is no communication about it.
Scaleway has terrible hardware and is not reliable at all. You can turn off an instance and not be able to turn it on again because they've run out of resources. IPv6 addresses change every time your instance is stopped and started.
Online.net owns scaleway so I assume their level of service is similar but with better hardware.
Re:Updated Linode Plans / New Larger Linodes
by devereaux on Thursday May 31, 2018 @02:41PM
(#17198897)
My experience with online.net/scaleway from just yesterday: their peering is quite bad, and their IPv6 is broken out of the box.
First, you only get a /127. Second, IPv6 is not activated by default, and for a good reason: when you do activate it, basic things like a apt-get update of a their vanilla distribution are broken.
Basically, their IPv6 peering seems so bad you can't even ping6 some ASN.
Practical example: on their Amsterdam DC, I couldn't even reach ftp.nl.debian.org (and before you ask, DNS worked just fine - I got the correct AAAA records)
When I asked for explanations and provided them a detailed ping6 and traceroute analysis showing where the problem was (and the same analysis from another server I had nearby that confirmed the problem was on online.net end), I was told to try the "rescue mode".
Re:Updated Linode Plans / New Larger Linodes
by LinuxBender on Thursday May 31, 2018 @02:07PM
(#17198586)
I've had numerous VM's with Linode for several years. I am also on DO and Vultr. The only reboots I get are for hardware failures and the recent Spectre patches. I would say on average, 1 reboot every 2 years. I could be lucky though.
Re:Updated Linode Plans / New Larger Linodes
by jethro_tell on Thursday May 31, 2018 @02:14PM
(#17198654)
Yes, That's what sticks out to me, time to patch, 6-9 months with a know, un-patched, vulnerability with no notice to customers. I wouldn't run a honey pot on Linode, their security record is abysmal and that's only the stuff we've found out about.
I'm stuck on Vultr still, since they're cheap and they have an Australian datacenter.
AU routing to JP is really bad (I get faster connections and better ping to TX-US servers on a lot of providers), and SG is like 120-160 ping which still kinda sucks.
Scaleway has terrible hardware and is not reliable at all. You can turn off an instance and not be able to turn it on again because they've run out of resources. IPv6 addresses change every time your instance is stopped and started.
Online.net owns scaleway so I assume their level of service is similar but with better hardware.
First, you only get a /127. Second, IPv6 is not activated by default, and for a good reason: when you do activate it, basic things like a apt-get update of a their vanilla distribution are broken.
Basically, their IPv6 peering seems so bad you can't even ping6 some ASN.
Practical example: on their Amsterdam DC, I couldn't even reach ftp.nl.debian.org (and before you ask, DNS worked just fine - I got the correct AAAA records)
When I asked for explanations and provided them a detailed ping6 and traceroute analysis showing where the problem was (and the same analysis from another server I had nearby that confirmed the problem was on online.net end), I was told to try the "rescue mode".
Lol no. I tried another company instead.
I've had a handful of nodes on their $20 tier for 10+ years, and I think I've had two or three unannounced reboots in that time.