Hacker News Reader

Comment by mcescalante | Hacker News Reader

Does anybody have a good comparison or story of why Linode is "more reliable" than DO? I've read the simple "Digital Ocean vs. Linode" stuff posted here, but it never offers any quantitative insight as to why people find Linode more reliable. Anecdotes encouraged :)

View on HN

I've had five 1GB VPSes running on Digital Ocean for the past year, for http://www.hostedapachesolr.com/. Of the five of them, three seem to have 5-10 minute downtimes about once per week as measured by https://servercheck.in/.

On Linode, I stopped using their 2 GB VPSes for the past year, but before that, I had 3 of them. One had about 5 minutes of unscheduled downtime over the course of a year, the other two only had a few scheduled restarts where Linode emailed me a day or so before it happened.

Anecdotal evidence, to be sure, and I've been extremely happy with both providers (Digital Ocean has been very good, much better than any other sub-$20/month provider until today's news)—but when push comes to shove, I would rather have my more critical VPSes on Linode.

I can only give you anecdotal evidence on two Linodes we use:

  root@web1:~# uname -a
  Linux web1.netstock.co 3.9.3-x86-linode52 #1 SMP Mon May 20 09:32:28 EDT 2013 i686 i686 i386 GNU/Linux
  root@web1:~# uptime
   15:29:46 up 328 days, 13:16,  4 users,  load average: 0.30, 0.15, 0.12

and

  root@li346-160:~# uname -a
  Linux li346-160 3.7.5-linode48 #1 SMP Thu Jan 31 14:40:54 EST 2013 i686 i686 i386 GNU/Linux
  root@li346-160:~# uptime
   15:31:11 up 487 days, 11:10,  6 users,  load average: 0.17, 0.09, 0.07

My server had 1000+ days uptime before it had to be restarted (migration/upgrade).

Do you never patch the kernel for security updates?

Every time I hear/see these crazy high uptime entries I cringe a little at the security implications. Unless you're applying updates to everything but the kernel, but still there have been at least 2 local vulnerabilities in the past few years.

DO will take your site offline without even telling you the reason why. Have done it to me multiple times, and only tell me after I file a support ticket. They seem to be disorganized and make decisions separately/randomly.

http://status.linode.com/

vs.

https://status.digitalocean.com/

I don't trust people's opinions unless either of them get caught outright lying on their status pages. ;)

Anecdotal: I've seen ~5x as many VPS restarts due to host issues on DO than Linode.

DO pretty much never seems to post anything on their status page unless it affects a lot of their customers -- I've had support tickets with them where they say a problem is affecting "some" customers, but they pretty much never posted it on their status page. Whereas Linode seems to post status problems far more frequently, and pretty much none of the problems listed have ever affected any of my nodes.

Of course, this is all anecdotal, but Linode really does seem to over report on their status page. And DO perhaps under reports.

This. My experience has been that Linode is very transparent about status messages, and I'd even go so far as to say the most transparent I've seen. Usually if an external service reports a timeout to me and I visit the status page there is often an investigation notice posted regarding network latency.

to be fair Linode also only posts when outages are DC-wide or otherwise affect large numbers of users.

I've saw restarts and networks issues frequently on DO months ago. They seem to have stabilized in the NYC datacenters where I host https://minimalreader.com/ but overall, the stability always seems to be on the edge of collapse. I'm not sure if that's because they are growing so quickly they cannot keep their network and hosts up with the demand, or what.

Not a useful metric unless both use absolutely identical criteria for posting.

I have only good experience with both. My linode (internet exposed) server has been running for almost 7 years now. My DO server (also internet exposed) is ~2 years old. These days I would spin up a DO by default because I am sucker for SSD. I might switch back to defaulting to linode with this new pricing. linode has a fantastic irc channel. Plus linode has IPv6 support which DO has promised _years_ ago. Just hop on and the linode guys help you out immediately.

I don't think you can go wrong with either for medium sized companies (< 100).

Any newly created Linode will land on a host with full SSD storage.

Linode recently converted to SSD as well.

pingdom reports a 99.98% uptime for my DO box (8min of downtime in the last month). To a commercial user that may be unacceptable, but it's great for a $5 box for me :)

Been running on DO for a while now without any downtime:

18:39:03 up 249 days, 4:57, 1 user, load average: 0.18, 0.20, 0.24

Moved off of Linode since we were down fairly often (to be fair it was Linode's Fremont DC) but we'd also have our dev server go down every month or 2 in Dallas.

Honestly I think it's probably hit and miss and depends a bit on luck and the server your VPS gets setup on.

DO's network speeds and latency are not very consistent for me. I can make a ticket and 'noisy neighbors' are moved to another node but the problems always come back.

Thank you guys for the replies, this is the stuff I was looking for :)

Reply on news.ycombinator.com