New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
DO launches yet another API (metadata) as they call it.
VPSRAIDSolutions
Member
in General
https://developers.digitalocean.com/metadata/
Link to the developer's documentation. Just felt like posting it since I had nothing much to do other than a CMS project currently.
Comments
Seems to be a clone of the Amazon metadata or the Openstack metadata: http://docs.openstack.org/admin-guide-cloud/content/section_metadata-service.html
Might be handy...
DO sucks
Is there a particular reason for that or you just jealous of their success?
+1 for the question
Just to name a few. At least they corrected some... details like using RAID5 on the nodes. But they still offer a uncertain, venture-backed service. Sorry but I would prefer a real company with bigger experience in the field, specially when they don't compete in pricing anymore.
I don't look at DO as a production environment however as a quick development environment i find it hard to beat.
(not saying it cant be used for production, just a personal preference)
Fair enough, if those are things you have issue with, most are not going to a concern to the average and majority of users.
You can't really hit them over the head for things like additional ip's though, it's not that they can't they simply don't offer it.
Yep, they are fine for personal use: cheap, kinda reliable and many locations around the world.
Main reason for using for development are take the cheapest droplet at $5 p/m thats $60 a year and can get a much better deal for that for production, however i dont mind paying a premium to have the facilities they offer to spin up and down as needed.
no custom kernel* (there are workarounds to make their kernel to load other but that is not near stable for production)
I have a good feeling that @linuxthefish said that because DO exempted his lot credits due to "abuse" of multiple coupon codes with github education pack.
Right word is few, not many iirc.
Nice list for the time they got in the market. South America and Australia are coming too, so I think it's pretty neat for the prices they got.
I'm just not a big fan of DO, sorry for going off topic and venting my anger! Network seems to be kinda crap, lots of drop outs when i tried to use for my BNC and disk speed does not seem anywhere near SSD...
10 requests completed in 9006.9 ms, 2407 iops, 9.4 mb/s
Compared with my GVH and Ramnode boxes...
10 requests completed in 9.0 s, 7.8 k iops, 30.6 MiB/s
10 requests completed in 9.0 s, 13.0 k iops, 50.7 MiB/s
What's up with this fake "cloud" crap anyway?
WAIT A SECOND. THEY'RE GOING TO GET AUSTRALIA?!
Just to show that its a similar (really even worse picture) with DO's competitor Vultr here is some from them -
10 requests completed in 9.0 s, 1.1 k iops, 4.3 MiB/s
You really do get what you pay for. Seriously how can they get iops that low?
Edit: Just to compare this to one of the online.net '1.99' servers -
10 requests completed in 9.0 s, 3.0 k iops, 11.8 MiB/s
If they are SSD's on DO and Vultr they are doing something messed up to destroy the performance so much
If you think you are getting 3K IOPS on a single 7200 RPM SATA HDD, you should apply for a Nobel prize.
That's what ioping reports for the 2 I have. Remember those little boxes, probably are very optimised on the IO front. Its just the CPU front where they are god damn awful. For instance the 1 cpu instance vultr servers run about 3 times as fast on the same CPU tests, while with another (even cheaper than Vultr) VPS I get 10 times the CPU performance compared to the 1.99 server.
Vultr and DO really are a lot of marketing, very little performance, but the ability to spin up servers as needed is a god send.
Where is my prize?
10 requests completed in 9005.3 ms, 2713 iops, 10.6 mb/s
Vultr 768Mb Instance:
[root@sogeek ~]# ioping .
4.0 KiB from . (ext4 /dev/vda1): request=1 time=193 us
4.0 KiB from . (ext4 /dev/vda1): request=2 time=351 us
4.0 KiB from . (ext4 /dev/vda1): request=3 time=308 us
4.0 KiB from . (ext4 /dev/vda1): request=4 time=380 us
4.0 KiB from . (ext4 /dev/vda1): request=5 time=323 us
^C
--- . (ext4 /dev/vda1) ioping statistics ---
5 requests completed in 4.1 s, 3.2 k iops, 12.6 MiB/s
min/avg/max/mdev = 193 us / 311 us / 380 us / 63 us
[root@sogeek ~]# w
07:12:47 up 169 days, 15:22, 1 user, load average: 0.03, 0.01, 0.00
[edited for formatting]
Try it without caching. 7k disks do only have around 100 IOPSs.
Without caching was not a requirement :P