Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
hurrah... :-|
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
mrbassie
l33t
l33t


Joined: 31 May 2013
Posts: 601

PostPosted: Tue Jun 20, 2017 9:58 pm    Post subject: hurrah... :-| Reply with quote

https://phoronix.com/scan.php?page=news_item&px=Lennart-casync
Back to top
View user's profile Send private message
Ant P.
Watchman
Watchman


Joined: 18 Apr 2009
Posts: 6618

PostPosted: Wed Jun 21, 2017 12:15 am    Post subject: Reply with quote

Non-clickfarm link: http://0pointer.net/blog/casync-a-tool-for-distributing-file-system-images.html

Just another redhat solution to a problem caused by a previous redhat solution to a problem caused by ...
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 45792
Location: 56N 3W

PostPosted: Wed Jun 21, 2017 9:00 am    Post subject: Reply with quote

Can I put it in package.mask already?
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
depontius
Advocate
Advocate


Joined: 05 May 2004
Posts: 3426

PostPosted: Wed Jun 21, 2017 3:53 pm    Post subject: Reply with quote

NeddySeagoon wrote:
Can I put it in package.mask already?


I'd be careful about that move. No doubt a future version of fvwm will be made to depend on it. Or pick your favorite WM, or any other package that should have nothing to do with an rsync-alike.
_________________
.sigs waste space and bandwidth
Back to top
View user's profile Send private message
R0b0t1
Apprentice
Apprentice


Joined: 05 Jun 2008
Posts: 255

PostPosted: Wed Jun 21, 2017 5:08 pm    Post subject: Reply with quote

That picture has a lot of lines in it! This must be real programming. I don't think I understand it. Maybe some day I will be smart enough to understand complex software projects like casync. I think for now rsync or borg backup is enough, but maybe I am thinking about filesytem images differently than the author.
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7427

PostPosted: Wed Jun 21, 2017 6:57 pm    Post subject: Reply with quote

R0b0t1 wrote:
I don't think I understand it. Maybe some day I will be smart enough to understand complex software projects like casync.

It's quiet easy, suppose you want to program a wheel

1/ blame current wheels, because WTF the wheels are all rounded, what a past technology to use rounded wheels
2/ reinvent the wheel, of course you'll do a rounded one too, but who cares about step1 anyway! Just make sure your wheel cannot be use by other wheel users ; it's important that your wheel could only be use by users able to only use your kind of wheel, because as soon as they switch to use your wheel, even if they see your wheel is totally crap, they couldn't afford to get back to good old wheels that just works.
3/ promise everyone next features will blown their mind! self powered wheel, auto repair itself, some floaters in case someone wants use it on water...


sure step 3 could bug you, but you're not insane and you'll just ignore step 3, they have switch to your wheel because of step 3 promises, no need to do step 3, they already use the wheel.

you may ask, what is step 4 then?
step4 is picking up another program, let's say this time you'll reinvent well, the horse....

and....

Ok, the magic is here, i hope you're sit!
...and make that new horse using your wheels... you're now a genius!!!
Don't worry if some complain it's stupid to have an horse with wheels, just make sure you'll blame enough in step1 other horses that cannot use any wheels.
Back to top
View user's profile Send private message
szatox
Veteran
Veteran


Joined: 27 Aug 2013
Posts: 1896

PostPosted: Wed Jun 21, 2017 11:22 pm    Post subject: Reply with quote

LOL, krinn, good one
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Thu Jun 22, 2017 3:12 pm    Post subject: Reply with quote

Genius, krinn :-)

There's nothing wrong with exploring an area, and writing a program to do what you want, specifically.
There's also nothing wrong with exploring an unfamiliar idea, or area of programming.

From the blog post, I don't see why squashfs images, and zsync for updates, do not fulfil the primary use-case of filesystem images.
Each use-case can use the tools best suited to it.
(HTTP ranges are hardly a major requirement, nowadays; using the possible lack as rationale to reinvent a tool makes little sense, especially when you pretty much require an sshd.)

Personally, I am glad Poeterring has something else to occupy his mind; there's certainly less scope for a Master Control Program here.
No doubt he thinks he's already got that in systemdbust.

The bit about "no random access in tar" made me smile, as cpio is nice for that, and standardised by POSIX ages ago.

As usual, it feels like curiousity with an idea combined with partial knowledge, turning into enthusiasm for all the ways it could be applied, becoming "it must be used for" those because of feature X that is only (potentially) relevant to another use-case.
Meanwhile, here's a "technical rationale" that reads more like marketing, caught up as it is with that enthusiasm and the laundry-list of "features".

That's not to say the "content-addressable", or more simply "hashed" part isn't interesting; but it would be more useful in terms of deduplication, and that would require a consistent chunk-size, say 4K.

Still, I don't have an issue with Poeterring scratching his itch.
I'm just glad he's in a smaller sandbox this time.
Back to top
View user's profile Send private message
miket
Guru
Guru


Joined: 28 Apr 2007
Posts: 424
Location: Gainesville, FL, USA

PostPosted: Sat Jun 24, 2017 3:07 am    Post subject: Reply with quote

steveL wrote:
Still, I don't have an issue with Poeterring scratching his itch.
I'm just glad he's in a smaller sandbox this time.

Really? Are you sure about that? I'm thinking that at some point he'll decide that it's just no good to distribute system images with having an init system to start them.

Or maybe the other way around. Here is a nice, handy system-image subsystem that could be integrated somehow into an existing init or system-management framework. Or something like that.
Back to top
View user's profile Send private message
devilheart
l33t
l33t


Joined: 17 Mar 2005
Posts: 837
Location: Villach, Austria

PostPosted: Sat Jun 24, 2017 6:55 am    Post subject: Reply with quote

I never had to distribute filesystem images, but I still fail to see which problem is this tool trying to solve
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sat Jun 24, 2017 4:32 pm    Post subject: Reply with quote

steveL wrote:
Still, I don't have an issue with Poeterring scratching his itch.
I'm just glad he's in a smaller sandbox this time.
miket wrote:
Really? Are you sure about that? I'm thinking that at some point he'll decide that it's just no good to distribute system images with having an init system to start them.
Sure, we know the kind of "system" he wants to distribute; as I said, he's already got a Master Control Process (with all its inherent failings.)
Given that, any other tool doesn't need to do the spaghetti-tentacle thing, and so he's back to just working on the tool itself.

The reality of having to deal with a new domain, means he's naturally going to (start to) appreciate modularity: no need to do anything but the tool itself, and worrying about things that are outside the domain is self-evidently a waste of headspace.

Of course, it's already vague in its aims, since it is driven by exploring an idea (using hashes) and not so much by the specific motivating application (distribution of file-system images), which would be better done with squashfs and zsync, according to his own post.
That's fine: it's his time, and his itch.

Greybeards tutting about how badly his lack of UNIX understanding (whereby wrapping other tools is the first option, not the last) shows he actually needs to spend at least a year learning shell, isn't going to affect him.

IME, the reality of programming means you eventually come to relish the idea of something being Somebody Else's Problem (where once you wanted control over, or credit for, everything.)
That's what people are usually getting at when they use the word "orthogonal": it's not relevant to this scope (so we don't have to worry about it.)

You're right though; it's no doubt aimed at being part of the whole RedHat GnomeOS logjam.
Certainly gets round the whole distribution package-management process, now politicized. No need to enforce the "gentle Putsch" via spurious dependencies; having poisoned the collaborative well for everyone else, so that most distros are closer and closer to "a derivative of RedHat" (and thus losing more and more of their "product differentiation", in business terms), the next phase seems to be "just use one of our images" for your next VM.
I mean, "everyone else is just downstream of RedHat anyway", so why not get the original (with paid support, and secret fixes.) You won't get fired for it..

This might all just be because of Oracle, or also just the nature of business under crapitalism, whatever. The end result is the same: corporations seek to coopt the work, by closing down collaboration amongst the technically-minded who crafted the product in the first place, and continue to maintain both the work, and the infrastructure.
That's what the GPL was all about; and of course the tendency hasn't gone away. Naturally corporations seek other means to put up barriers to entry (like "support contracts" that are unilaterally terminated should you *shock horror* exercise your GPL rights.)

"Commercial distribution is killing Free software."
Back to top
View user's profile Send private message
nicop06
n00b
n00b


Joined: 25 May 2017
Posts: 13
Location: France

PostPosted: Sun Jun 25, 2017 6:06 pm    Post subject: Reply with quote

After reading the blog post I am a bit confused. Isn't casync trying to solve the same problem that BTRFS and ZFS snapshots solved years ago? At loast for the backup part.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum