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.
Comments
I may need to wait until they replace the only drive in the server that is available to deploy on. Could be a couple of hours, pretty inconvenient.
Unless you have a fancy automatic recovery system that recreates the VM, you're still going to need manual intervention to rebuild it, and whether it takes you a couple of minutes or a couple of hours the application is still down until you intervene.
Whether or not there are known case, transparency enables customer to make more informed decisions. Also it's hard to measure the reliability since they don't publicly publish problems like this as far as I know, unlike RamNode and some others.
Personally I don't mind since my application isn't critical, but if I were running a service that was generating much more money, I would want to know more details about the infrastructure. iwstack.com is particularly transparent, and AWS / Google Compute / etc. already have solid reputation that Vultr doesn't have (no idea how transparent they are though).
Edit: well I guess if it was generating much more money then I ought to have an automatic failover solution and templates to provision new nodes from. Maybe a bit more money, but not so much that I'd want to spend time on that? lol. Anyway point is I could go with provider who I can better gauge reliability without any other disadvantages. Also it's not just about whether they have RAID, but also the attitude that you are able to see from them, there are some people who can give you solid deal but maybe you don't want to support them.
Heh, why not KISS and just rebuild from a snapshot in under 10 minutes, all told? Better yet, if it's important enough, have a cloned shadow idling in another location
KAGE BUSHIN NO JUTSU!
風遁・螺旋手裏剣
Nice try,.. Vultr? no thanks, I will stick with DO.
Two weeks ago that gif would have ruined my life. Yes, I am that far behind on Naruto.
I know what you mean... I started watching Naruto from Episode one in November 2013... I'm all caught up now
What episode are you at right now?
Same, I'm sure at least one person here would have lost all of their data by now, unless Vultr have magic drives that never fail.
238, slap bang in the filler arc ahead of Naruto meeting Bee for the first time.
I've had to have a break, so watching other stuff for a week or two before I dive back in.
Yep. It's all common sense. Can you imagine all these hypervisors running 24x7 without a single drive failure for over a year now? I wouldn't it. If someone lost their data, they would have gone public about it, especially in a community like LET.
While I do not fully agree with David on his decision to simply not disclose his hardware setup, I do respect it. Why? Simply because there's no written rule that obligates him to reveal all the details... It's a matter of choice, that's all. It may upset some, but others won't care.
Let's just keep one thing in mind, Choopa has been in business for 15 years, they host one of EFnet's oldest and most reliable servers, they run GameServers.com etc... I highly doubt that a company of their caliber would half-a$$ their setup and not use proper raid and take proper measures for redundancy.
Furthermore, do you really think that the I/O you get on their VM's can be achieved and sustained without RAID?
Good GOD... You're SO BEHIND! I am so tempted to post some spoilers
Is it for old member
Generally you should operate under the assumption that whatever "cloud" or VPS provider you use, will lose your VM later today, ban your account by the evening and go out of business tomorrow. None of those should present any significant issue for you whatsoever.
Sure some manual intervention will be needed, but if your app is truly critical, you should design it as a fault-tolerant cluster running at 2-3 providers with real-time synchronization (using a distributed filesystem, SQL replication, etc).
I was level up until that awful filler arc with the kid and the woman who made crystals, but that was so bad I stopped watching for a couple of years.
In the absence of official RAID1/5/6/10 confirmation, I shall continue to treat vultr instances like AWS ec2's old ephemeral storage instances(but with raid0 ssd performance).
No non-replicated mysql/rdbms running here
My $50(x2=100) promo deposit is still working fine for running experiments for a handful of hours with whatever newfangled distributed software/databases that catch my fancy (a new one announced every hour?)
Heck, maybe vultr will teach us all to design systems to fail fast and recover even faster.
(And on a low-end budget)
Given the state of things, I would welcome an API flag/call that tells the vultr internal hypervisor selector to ensure that that '/v1/server/create' allocates a new instance from a 'fresh' hypervisor node . (where 'fresh' = none of my already running instances are hosted)
At the very least, this should be statistically feasible in one of their major locations -NJ/Dallas (things get iffy if they are bundling only instanceX types onto the same hypervisor hardware nodes)
Oh yeah, Guren.... Well, it wasn't a filler, man. It's part of getting the 3-tails. Have you figured out who Tobi is yet :P
OVH now offer servers where you can get 4 x SSD/SAS optionally for additional charges
e.g http://www.ovh.ie/dedicated_servers/big-data/mini-HG.xml
But yes this was not the case until a few months ago
@fametel you could also just get a storage server from them they have more than two disks.
Generally OVH only offers two disks.
Yes, I know that but they released their storage line in 2014 for the first time, before that the only way of getting more than 2 disks was ordering their HG servers. And their storage servers are not suitable for general vps offerings due to the low amount of RAM unless you want to offer a storage line like backupsy or cloudshards.
You can now get 3 disks with OVH's general Enterprise, Hosting or Infrastructure rangers also by just paying additional 5 Euros per month.
It was filler apart from Akatsuki getting the three tails. The rest is worthless.
Yeah, I know who Tobi actually is.
It's been a long time since I've seen that episode... I suppose by now you'd know who Tobi is.. I am against resisting spoiling it for you
Edit: Best derailing ever!
That's it. I don't want to have to manually intervene if a single drive fails, that's all. I doubt that's too much to ask.
I know how it all ends as I once I got behind I decided it wait until the manga finished to make sure the ending wasn't stupid before committing more time to the anime. I just don't know how it gets to the epilogue/Naruto the Last movie.
ooh