5 ms·

This answer doesn't include any mention of OpenVZ, which means you're not asking a Linux expert. The answer sounds like the type confused competitors invariably

by staunch 11y ago

This answer doesn't include any mention of OpenVZ, which means you're not asking a Linux expert. The answer sounds like the type confused competitors invariably give, which boils down to "we did it first, they just made it flashy and cool" and is never the real reason. > When you think about why would someone pick one Linux distro over the other, it’s because of — literally — the aesthetics of how the directories are laid out and what they’re called; the default shell and how much stuff was in there; and the packaging system. That was it. This is straight from a Slashdot flamewar in 1998.

Pyxl101 11y ago
What would the Linux expert say about OpenVZ?
- staunch 11y ago
  At the very least, that Linux-based OpenVZ containers have been widely deployed for many years in many orders of magnitude higher volumes than FreeBSD Jails, Solaris Zones, or anything else.
dasil003 11y ago
It's worse than that. I was there at the founding of TextDrive, I shelled out the cash for "lifetime hosting" which got them off the ground. You can call this sour grapes, because it definitely is, but this picture that Jason paints about how awesome his technology was and how he was there first on all this stuff is laughable—because the technology did not work. When he says something like: > ...the smartest thing we could’ve done was never let all the web 2.0 companies go away. We had everybody back then. It was probably the biggest tactical error, if you will, but the tech was fine. The reason they lost those customers was that their tech was not fine. In the early days, Jason was posting dozens of posts a day to the TextDrive forum waxing on about how awesome the tech they were working on was, however at no point did it ever stabilize into something reliable. It was initially sold as being the first non-oversold shared hosting service, which implicitly promises stability, but in fact it was half the reliability of traditional shared hosting like Dreamhost at 10x the price. Now it was true that they were bleeding edge, and so they got a lot of these companies on board, and even were the first "official Rails host" before any other shared hosting could handle Rails, and they had a lot of brand equity, which is how they managed to have this great early customer list. But it was all just based on hype and a compelling sales pitch from Jason. It really rubs me the wrong way how he has parlayed his mediocre-at-best technology execution into an image as a technology pioneer which he continues to use to leverage himself to new heights. The guy is a mediocre technologist, but a fantastic salesman.
- bsaul 11y ago
  I wasn't a textdrive user and actually didn't even know about it, but just reading that part of the interview made me think the tech wasn't that fine. he goes from "people moved to amazon because of s3 which worked so well" to "our tech was fine". Seems like he considers s3 a marketing trick..
- jimpick 11y ago
  Ex-Joyent employee here, but I don't go back that far. The company has quite the history. I have to disagree with your characterizations. The products changed much over the years, many things were tried, many failed. Most startups change their product significantly over time and it's hardly surprising that the business they are in now isn't one of the early iterations they tried and failed at. I strongly believe that Jason is one of the best technologists on the planet. So there.
  - wpietri 11y ago
    What puzzles me about your comment is that you don't directly refute what Dasil003 was saying, which is that the technology didn't actually work, that from the customer perspective the execution was "mediocre at best". I'm sure he's a bright and likable guy, and from his resume I figure he must be an inspiring leader. But for me, being a great technologist is all about things actually working, about users actually getting served. Sure, one should fail a lot in the lab, and sure, startups should be trying enough things that some just don't have product/market fit. But that's very different to me than promising the moon and delivering stuff that doesn't actually work.
    - jimpick 11y ago
      During my time there, the stuff was amazingly solid. It was Solaris, not Linux, with a heavy dose of NetBSD ports. The customer support was the best in the biz too, but only the larger customers really saw that. There were always those legacy customers from years before who had there stuff running on some boxes in an old datacenter that only a few of the original staff knew about, and required continuous care and attention. Ihere's a business lesson to learn there - if you sell a "lifetime" subscription for a product that will become obsolete, you'll pay a lifetime PR cost in the future when you can't continue with that product. Coming from a pretty long Linux background, Solaris took some getting used to, but once I figured it out, it really was a better, more solid, better engineered solution. But Linux has a huge community - the people that loved Solaris are the old Unix graybeards, and that community wasn't growing. Before my time, there were some debacles, such as a storage platform product that ended up losing data, but by the time I was there, they had learned the lessons, and I'll still say their core stack is still superior to any Linux based stack used in the datacenter today. It's been 4 years since I worked there.
tywkeene 11y ago
From my perspective (which is rather limited, I'll admit) this actually does seem to be why docker is so popular. Could you elaborate on why this isn't the reason docker is so big now?
vidarh 11y ago
He might not be a Linux expert, but as someone who has admin'ed Linux systems since '94, and has used OpenVz in production since 2006, and LXC since 2007, I agree with his assessment of why people are flocking to Docker. As a longtime OpenVz user it's a given that we have to migrate off it at some point. Until Docker, LXC or KVM were the viable alternatives for me, and neither appealed: LXC seemed too half baked, though the basic elements were there; KVM too heavy for our use. Docker did appeal to me exactly because it provided a simple story around how to package up and distribute container images that made LXC simple (before they provided their own backend) and let us continue with containers instead of going the VM route. With respect to the quote you give, it somewhat exaggerates how superficial the differences are, but not all that much: It is largely about how files are laid out and package management. But those are huge things because they pretty much define the whole system and how we interact with them. Note how many projects out there are even trying to find ways of using Docker and similar to package user-level applications too...
- strictfp 11y ago
  Have you considered user-mode Linux (UML) as something in between LXC and KVM?
  - vidarh 11y ago
    I haven't considered UML for anything for many years. I can't really see any advantages to UML these days.
timbutlerau 11y ago
Virtuozzo / OpenVZ brought container technology to hosting providers and according to Odin has now had over 1 million containers deployed. We use them for our company and wouldn't be able to achieve the reliability nor the performance without them. We're also excited about Docker, because it brings a whole new set of tools to help complete the whole picture. The container system is just one small part of Docker, it's the whole toolset and platform which is why it's so popular. The best thing is, Docker will run in nested containers with Virtuozzo / OpenVZ, so you can take direct advantage of both technologies.
- digitalzombie 11y ago
  > We're also excited about Docker, because it brings a whole new set of tools to help complete the whole picture. >___> there's never going to be a complete picture. There will always be a new shiny toy out there to play with and to go with that complete picture. Google probably have other shiny toys that complement those containers that they haven't talked about yet. They had borg for a long time and even now we're barely getting some of those, Mesos & Kubernates.

by staunch 11y ago

Pyxl101 11y ago

What would the Linux expert say about OpenVZ?

staunch 11y ago
At the very least, that Linux-based OpenVZ containers have been widely deployed for many years in many orders of magnitude higher volumes than FreeBSD Jails, Solaris Zones, or anything else.

dasil003 11y ago

It's worse than that. I was there at the founding of TextDrive, I shelled out the cash for "lifetime hosting" which got them off the ground. You can call this sour grapes, because it definitely is, but this picture that Jason paints about how awesome his technology was and how he was there first on all this stuff is laughable—because the technology did not work. When he says something like: > ...the smartest thing we could’ve done was never let all the web 2.0 companies go away. We had everybody back then. It was probably the biggest tactical error, if you will, but the tech was fine. The reason they lost those customers was that their tech was not fine. In the early days, Jason was posting dozens of posts a day to the TextDrive forum waxing on about how awesome the tech they were working on was, however at no point did it ever stabilize into something reliable. It was initially sold as being the first non-oversold shared hosting service, which implicitly promises stability, but in fact it was half the reliability of traditional shared hosting like Dreamhost at 10x the price. Now it was true that they were bleeding edge, and so they got a lot of these companies on board, and even were the first "official Rails host" before any other shared hosting could handle Rails, and they had a lot of brand equity, which is how they managed to have this great early customer list. But it was all just based on hype and a compelling sales pitch from Jason. It really rubs me the wrong way how he has parlayed his mediocre-at-best technology execution into an image as a technology pioneer which he continues to use to leverage himself to new heights. The guy is a mediocre technologist, but a fantastic salesman.

bsaul 11y ago
I wasn't a textdrive user and actually didn't even know about it, but just reading that part of the interview made me think the tech wasn't that fine. he goes from "people moved to amazon because of s3 which worked so well" to "our tech was fine". Seems like he considers s3 a marketing trick..
jimpick 11y ago
Ex-Joyent employee here, but I don't go back that far. The company has quite the history. I have to disagree with your characterizations. The products changed much over the years, many things were tried, many failed. Most startups change their product significantly over time and it's hardly surprising that the business they are in now isn't one of the early iterations they tried and failed at. I strongly believe that Jason is one of the best technologists on the planet. So there.
- wpietri 11y ago
  What puzzles me about your comment is that you don't directly refute what Dasil003 was saying, which is that the technology didn't actually work, that from the customer perspective the execution was "mediocre at best". I'm sure he's a bright and likable guy, and from his resume I figure he must be an inspiring leader. But for me, being a great technologist is all about things actually working, about users actually getting served. Sure, one should fail a lot in the lab, and sure, startups should be trying enough things that some just don't have product/market fit. But that's very different to me than promising the moon and delivering stuff that doesn't actually work.
  - jimpick 11y ago
    During my time there, the stuff was amazingly solid. It was Solaris, not Linux, with a heavy dose of NetBSD ports. The customer support was the best in the biz too, but only the larger customers really saw that. There were always those legacy customers from years before who had there stuff running on some boxes in an old datacenter that only a few of the original staff knew about, and required continuous care and attention. Ihere's a business lesson to learn there - if you sell a "lifetime" subscription for a product that will become obsolete, you'll pay a lifetime PR cost in the future when you can't continue with that product. Coming from a pretty long Linux background, Solaris took some getting used to, but once I figured it out, it really was a better, more solid, better engineered solution. But Linux has a huge community - the people that loved Solaris are the old Unix graybeards, and that community wasn't growing. Before my time, there were some debacles, such as a storage platform product that ended up losing data, but by the time I was there, they had learned the lessons, and I'll still say their core stack is still superior to any Linux based stack used in the datacenter today. It's been 4 years since I worked there.

tywkeene 11y ago

From my perspective (which is rather limited, I'll admit) this actually does seem to be why docker is so popular. Could you elaborate on why this isn't the reason docker is so big now?

vidarh 11y ago

He might not be a Linux expert, but as someone who has admin'ed Linux systems since '94, and has used OpenVz in production since 2006, and LXC since 2007, I agree with his assessment of why people are flocking to Docker. As a longtime OpenVz user it's a given that we have to migrate off it at some point. Until Docker, LXC or KVM were the viable alternatives for me, and neither appealed: LXC seemed too half baked, though the basic elements were there; KVM too heavy for our use. Docker did appeal to me exactly because it provided a simple story around how to package up and distribute container images that made LXC simple (before they provided their own backend) and let us continue with containers instead of going the VM route. With respect to the quote you give, it somewhat exaggerates how superficial the differences are, but not all that much: It is largely about how files are laid out and package management. But those are huge things because they pretty much define the whole system and how we interact with them. Note how many projects out there are even trying to find ways of using Docker and similar to package user-level applications too...

strictfp 11y ago
Have you considered user-mode Linux (UML) as something in between LXC and KVM?
- vidarh 11y ago
  I haven't considered UML for anything for many years. I can't really see any advantages to UML these days.

timbutlerau 11y ago

Virtuozzo / OpenVZ brought container technology to hosting providers and according to Odin has now had over 1 million containers deployed. We use them for our company and wouldn't be able to achieve the reliability nor the performance without them. We're also excited about Docker, because it brings a whole new set of tools to help complete the whole picture. The container system is just one small part of Docker, it's the whole toolset and platform which is why it's so popular. The best thing is, Docker will run in nested containers with Virtuozzo / OpenVZ, so you can take direct advantage of both technologies.

digitalzombie 11y ago
> We're also excited about Docker, because it brings a whole new set of tools to help complete the whole picture. >___> there's never going to be a complete picture. There will always be a new shiny toy out there to play with and to go with that complete picture. Google probably have other shiny toys that complement those containers that they haven't talked about yet. They had borg for a long time and even now we're barely getting some of those, Mesos & Kubernates.