If I read this correctly, these components run on top of open source fork of Solaris, SmartOS?
Or SmartOS is using parts of Illumos, another for of open source Solaris, but it's not really a Solaris with its userland, but a container OS? Like Linux + Docker, but without the OS parts of Linux distribution.
The whole thing sounds very intriguing and really useful for a lot of things. I would like to do a Python Pandas DataFrame on top of Manta objects, for instance.
Hopefully someone could get a Vagrant box or something to try the whole stack on your own hardware. What I'm looking at now, I don't know where would I start building. Other option is to use their cloud services, but looking at their price charts [1] I get overwhelmed with options (and no free tier to take it for a spin).
[edit: Jonathan Perkin answred this question much better that I could].
SmartOS is an illumos distribution tailored for running containers and KVM guests. It usually PXE-boots, but it can boot from things like USB sticks too. It's a "live" distribution in the sense that that there is (almost) no local state and no local configuration. It just boots and takes care of the containers or KVM guests living on the ZFS filesystems on that particular node.
SmartOS is trivial to get running, you just dd an image to an USB stick. You can create your own images to run, or you can import the many images provided by Joyent or the community. The tooling is very nice and mature.
SDC is additional technology that uses SmartOS (among other things) that you want to use when you want to run not a single node, but clusters of many nodes.
I have run this stack on my laptop, it requires a lot of ram (think 8GB+) and is not easy to set up, especially Manta, even when setting up from binaries. Maybe the setup part has become easier during the year I haven't touched it though.
* Between Solaris 10 and Solaris 11, Oracle bought Sun and killed the open source efforts known as OpenSolaris. illumos started from the final open source bits of what eventually became Solaris 11 and has now significantly diverged. Solaris is effectively dead, illumos is very much alive.
* There are a number of illumos "forks", of which SmartOS is ours, but all forks still share the common illumos code (we merge daily) and contribute heavily back to the common base. Each fork may contain features which aren't yet ready for merging back, e.g. our work to port KVM[1] to SmartOS is not part of illumos yet, but other distributions such as OmniOS[2] have taken that work and integrated it.
* SmartOS is a minimal distribution, we have removed a lot of parts (desktop, shared storage, etc.) which do not fit in with our explicit design goals, and added tooling around virtualisation. You boot from USB/CD/PXE into a minimal live-image hypervisor known as the Global Zone[3], and then perform work in zones which are backed by local storage. To upgrade, you simply replace the USB image with a newer platform and reboot into the new live image.
* As for software, we provide userland built from pkgsrc[4], which gives you access to over 13,000 packages available under /opt/local, allowing you to use both the SmartOS tools as well as any third party software you may need (e.g. GNU stuff). There is even full desktop stuff provided, should you want to use it[5].
* In Manta, when running a job you are basically running in a zone with as many pkgsrc packages pre-installed as we can manage (currently nearly 9,000), so the chances are that the software you need is available. If not, you can easily build it yourself and the store it back in Manta to use later as an asset[6] for your jobs.
* There is definitely a free tier for Joyent zones (search for "free" on https://www.joyent.com/products/public-cloud/pricing), I thought there was also an additional free tier for Manta but I can't see it right now, however it will only cost a few cents to do some basic tests in Manta and get a feel for what it can do.
I primarily work on pkgsrc, but I know a number of other engineers will be reading this thread and can comment in more depth on SDC/Manta, so feel free to ask any more questions or pop onto #smartos or #manta on Freenode IRC.
Ok, thank you and others very much for your detailed info. I did use OpenSolaris for a brief time, but kind of lost track after Oracle came and took it.
In regards of the free tier, now I found it, but its mention is placed as the last line, after all the paid plans and the part about "Contact sales if your requirements exceed everything we've planned a price for" ;)
About Manta, I'd like to know about what kind of data objects it supports? Like having a matrix of time series data and making quick selections and sorts, near real time?
Docs talked about hierarchical storage and search, which sounds really useful, too, but what is search in this context?
* Log into a Manta zone using bbc.png as my input file. This creates a zone, and maps in my file which is stored on the same machine (you are always operating on the same host as your data is stored). mlogin(1) is a nice way to prototype jobs in an interactive session, and once you have it working correctly you can use mjob(1) to run it automatically.
$ mlogin /jperkin/public/tmp/bbc.png
* created interactive job -- f1a2e579-34f8-4dd4-da19-db33954a0772
* waiting for session... | established
jperkin@manta #
* At this point it's just Unix, so I can run any command on the file (which has been mapped in under /manta) I like:
jperkin@manta # uname -a
SunOS 0ae1c6ec-d47a-455c-9dd6-97eec16da31b 5.11 joyent_20140628T000418Z i86pc i386 i86pc Solaris
jperkin@manta # ls -l /manta/jperkin/public/tmp/bbc.png
-rw-r--r-- 1 root root 540751 Nov 7 11:41 /manta/jperkin/public/tmp/bbc.png
jperkin@manta # file /manta/jperkin/public/tmp/bbc.png
/manta/jperkin/public/tmp/bbc.png: PNG image data, 1680 x 940, 8-bit/color RGBA, non-interlaced
* Note that only the file I chose has been mapped in:
Of course this is a simple and contrived example, the real power of Manta comes when you have say 1,000,000 log files stored under a particular path and want to grep them all for a particular string. To do that you'd do something like:
This will scale to whatever size your Manta cluster is, e.g. if you have 10 hosts then the log files will be split up across those hosts and they each will spin up multiple zones to run "gzcat | grep" on the local data, before a final "cat" reduce job is used to collate the results from each map job.
Wow, thank you, again. I have to definitely take a look this. My use cases tend to vary so much that creating a Hadoop like system would require too much custom coding.
I wonder if it is possible to have compression and de-duplication, so that there could be a one big base dataset and lots of containers that only add what new data they generate.
Anyhow, looking at this it feels really approachable. What I have in mind are quick-and-dirty data-sciency scripts for ad hoc use cases, like diffing structured files and combing over matrix data.
I saw a Joyent manta demo at a node.js roadshow/meet-up a year or so ago, and it was impressive. It's lovely to see the action (code)/object (data) relationship rethought, whereby they said "let's bring the tools to the data instead of lugging data to the tools." It's also cool, refreshing, and validating to see "the tools" being Unix. "Imagine you want to grep across a dataset and sort those records and pull out some key fields with awk..." No need to imagine: that could be your exact plan, using the familiar tools you know and love.
I'm looking forward to trawling (ha! more nautical theming!) this code for the education, and maybe some gems.
If anyone's interested in a presentation introduction to Manta (presumably similar to the one you saw), you may want to check out my FutureStack presentation from last year.[1] And if you have 90 minutes to kill and are in the mood for the true origin story of Manta, see my video from NYC DevOps in January.[2]
Oh wow, my crush on Joyent just increased by two orders of magnitude. First, now (finally!) being open, manta is a viable alternative, and second, landing on MPL v2.0 seems like a very nice and pragmatic choice.
I highly recommend checking out the linked talk on Open Source anti-patterns for those, that like me, haven't seen it. I do think Bryan is a little quick on dismissing the GPL/CDDL problem (implying that RedHat didn't want dtrace) -- but I do see how the CDDL could evolve as unfortunately GPL incompatible, rather than intentionally so (he goes into this a little in the USENIX talk on OpenSolaris/Illumous).
This is cool. But, uh, how do I see the source to Manta? I read through all of the README on https://github.com/joyent/manta/, and cloned it, and can't figure it out.
Is there not a meta fetch/build tool like Chrome/Android have? I have built those two projects from scratch with no problems.
Or do you have to start with the sdc repo?
EDIT: I just cloned the repos manually. I thought that there would be some automated way to start from the manta repo and get all dependencies.
We are heavily using it in production on http://casinorpg.com (not a common use-case, but it works well for us). Basically, when a player creates a new character, they are able to customize out of millions of unique possibilities. Since the game is HTML5, we must generate a unique PNG sprite for their character. We run Manta jobs to do this by generating the image, compressing it and then storing it directly in Manta (where it then gets picked up by a CDN).
Given that you're asking from a throwaway (and especially given your choice of nickname), you're almost certainly trolling -- but given how old this thread has become, I think it's safe to answer here for the sake of posterity...
OpenStack and SmartDataCenter address some similar problems: they both manage fleets of physical machines and provide orchestration (provisioning, monitoring, etc.) of standing virtual machines on those physical machines. But in every other regard, these two projects are very dissimilar: they differ in their goals, in their organization and in their technology choices.
Getting more specific, I (obviously) have a dog in the fight, so it's hard for me to look at this with an outsider's perspective -- but I think the most nonjudgemental way of phrasing the fundamental difference is that SmartDataCenter is opinionated where OpenStack gives architectural choice to the integrator and/or operator.[1] So in SmartDataCenter you don't pick the storage substrate (it's ZFS) or the hypervisor (it's SmartOS) or the network virtualization (it's Crossbow). While OpenStack deliberately accommodates vendor differentiation, SmartDataCenter deliberately rejects it: we are designed for commodity storage (shared-nothing -- and no RAID controllers, please!), commodity network equipment (no vendor-specific SDN) and (certainly) commodity compute. The upshot is that the integrator/operator needn't design the system themselves -- which we know from experience can result in greatly reduced times of deployment. (Indeed, one of the great prides of SmartDataCenter is our install time: provided you're racked, stacked and cabled, you can get a cloud stood up in a matter of hours rather than days, weeks or longer.)
That fundamental difference -- opinion vs. choice -- has many ramifications. For example, we have no interest in governance (sorry, democrophiles!); if someone has a good idea, we'll do what most open source projects do and let our user community make that determination. This is not to be overly controlling; anyone who disagrees with us (or with anyone!) is welcome to fork the project -- we have deliberately selected a fork-friendly license in MPLv2. We also have no interest (zero, none, nada) in legacy enterprise hardware vendors that are interested in cloud computing only as a vector for preserving their inalienable right to screw their own customers; there won't be "hooks" or "plugins" in SmartDataCenter or generally other such sheep-like wolves.
There are many differences, of course, but I think many can be explained by the fundamental difference in engineering principles of the two projects.
Anyone who is interested in the nature of the opinions asserted in SmartDataCenter -- or in helping form new ones! -- should check out the repos[2][3] (we try to document our thinking), join the mailing lists, and/or join us in IRC at #smartos on Freenode.
[1] This is a bit of a "pro-choice/pro-life" nomenclature in that I have deliberately picked nomenclature that I think these two projects would use to put themselves in the best light. More candidly, I would say that OpenStack seems to me to be a discombobulated mess -- and I'm sure to OpenStack, SmartDataCenter would seem supremely fascist.
I think that that's completely fair -- and in fact, it's in the "Design principles" section of README.md in the SDC repo[1]: "SmartDataCenter is very opinionated about how to architect a cloud. These opinions are the result of many years of deploying and debugging the Joyent Cloud."
Or SmartOS is using parts of Illumos, another for of open source Solaris, but it's not really a Solaris with its userland, but a container OS? Like Linux + Docker, but without the OS parts of Linux distribution.
The whole thing sounds very intriguing and really useful for a lot of things. I would like to do a Python Pandas DataFrame on top of Manta objects, for instance.
Hopefully someone could get a Vagrant box or something to try the whole stack on your own hardware. What I'm looking at now, I don't know where would I start building. Other option is to use their cloud services, but looking at their price charts [1] I get overwhelmed with options (and no free tier to take it for a spin).
[1] https://www.joyent.com/products/public-cloud/pricing