Re: Re: [gentoo-dev] Idea for the portage maintainers

Re: Re: [gentoo-dev] Idea for the portage maintainers

[brettholcomb]

The twirl twirl is cache updating - it happens after any files that are synced scroll by and after it says updating cache.  Sync isn't the problem.  It's cache updating and that has been brought up in previous threads.  Maybe portage NG will address this.


> 
> From: Tom St Denis <tom@securescience.net>
> Date: 2004/04/12 Mon PM 05:21:56 GMT
> To: gentoo-dev@lists.gentoo.org
> Subject: Re: [gentoo-dev] Idea for the portage maintainers
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On April 12, 2004 01:19 pm, Norberto Bensa wrote:
> > Tom St Denis wrote:
> > > I started using gentoo [~60k files in portage] to now [~80k files in
> > > portage] it's definitely not just a linear amount slower.
> >
> > rsync isn't slow. "emerge --sync" is slow in the portage cache updates
> > stage (I really want to know what portage cache update does :-/ )
> 
> Syncs can be slow too.  Sometimes they zoom but sometimes it will say things 
> like
> 
> 200 files ......
> 
> and twirl..... and twirl......and twirl......
> 
> Point is portage is useable and works decent.  Updating is a bitch and 
> occasionally a "world" update takes way too long [to get the package list].  
> It's a manner of being scaleable.  Portage isn't.
> 
> Tom
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> 
> iD8DBQFAetA0sP+tEsHHY0ARAuJIAJ9jfAozHPtz2o49wiNQdni1FYU5FgCZAefS
> +qb63BEv19W73J6HN6v06WA=
> =CRvw
> -----END PGP SIGNATURE-----
> 
> --
> gentoo-dev@gentoo.org mailing list
> 
> 


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Andrew Gaffney]

brettholcomb@charter.net wrote:
 > The twirl twirl is cache updating - it happens after any files that are synced scroll 
by > and after it says updating cache.  Sync isn't the problem.  It's cache updating and 
that > has been brought up in previous threads.  Maybe portage NG will address this.

Can someone tell me exactly what is going on during a portage cache update? Or do I just 
need to Use The Source?

-- 
Andrew Gaffney
Network Administrator
Skyline Aeronautics, LLC.
636-357-1548


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 1151 bytes --]

On 04/12/04  Andrew Gaffney wrote:

> brettholcomb@charter.net wrote:
>  > The twirl twirl is cache updating - it happens after any files that
>  > are synced scroll 
> by > and after it says updating cache.  Sync isn't the problem.  It's
> cache updating and that > has been brought up in previous threads. 
> Maybe portage NG will address this.
> 
> Can someone tell me exactly what is going on during a portage cache
> update? Or do I just need to Use The Source?

Assuming default setup:
- copy /usr/portage/metadata/cache to /var/cache/edb/dep
- check mtimes of actual ebuilds (including PORTDIR_OVERLAY) and update
/var/cache/edb/dep if they are newer than their cache entries

So if you have a big PORTDIR_OVERLAY the cache update will take longer,
if you use RSYNC_EXCLUDEFROM it will also take longer.
The cache entries are the values of *DEPEND, IUSE, HOMEPAGE, ... in a
format that can be accessed without calling bash to parse the ebuild.

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 563 bytes --]

On 04/12/04  Marius Mauch wrote:

> So if you have a big PORTDIR_OVERLAY the cache update will take
> longer, if you use RSYNC_EXCLUDEFROM it will also take longer.
> The cache entries are the values of *DEPEND, IUSE, HOMEPAGE, ... in a
> format that can be accessed without calling bash to parse the ebuild.

Ignore my statement about RSYNC_EXCLUDEFROM, it's incorrect.

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Stuart Herbert]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

On Monday 12 April 2004 21:00, Marius Mauch wrote:
> Assuming default setup:
> - copy /usr/portage/metadata/cache to /var/cache/edb/dep
> - check mtimes of actual ebuilds (including PORTDIR_OVERLAY) and update
> /var/cache/edb/dep if they are newer than their cache entries

Will Portage function without /var/cache/edb?

Best regards,
Stu
-- 
Stuart Herbert                                              stuart@gentoo.org
Gentoo Developer                                       http://www.gentoo.org/
Missed the php|cruise?             http://dev.gentoo.org/~stuart/cruise-2004/

GnuGP key id# F9AFC57C available from http://pgp.mit.edu
Key fingerprint = 31FB 50D4 1F88 E227 F319  C549 0C2F 80BA F9AF C57C
--

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 604 bytes --]

On 04/12/04  Stuart Herbert wrote:

> On Monday 12 April 2004 21:00, Marius Mauch wrote:
> > Assuming default setup:
> > - copy /usr/portage/metadata/cache to /var/cache/edb/dep
> > - check mtimes of actual ebuilds (including PORTDIR_OVERLAY) and
> > update/var/cache/edb/dep if they are newer than their cache entries
> 
> Will Portage function without /var/cache/edb?

Yes, but it will be _very_ slow.

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Stuart Herbert]

[-- Attachment #1: Type: text/plain, Size: 759 bytes --]

On Monday 12 April 2004 21:58, Marius Mauch wrote:
> On 04/12/04  Stuart Herbert wrote:
> > Will Portage function without /var/cache/edb?
>
> Yes, but it will be _very_ slow.
>
> Marius

Then we shouldn't be putting this in /var/cache.  It's for temporary files, 
and can be emptied at any time without warning.  This should really go 
into /var/db.

Best regards,
Stu
-- 
Stuart Herbert                                              stuart@gentoo.org
Gentoo Developer                                       http://www.gentoo.org/
Missed the php|cruise?             http://dev.gentoo.org/~stuart/cruise-2004/

GnuGP key id# F9AFC57C available from http://pgp.mit.edu
Key fingerprint = 31FB 50D4 1F88 E227 F319  C549 0C2F 80BA F9AF C57C
--

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Spider]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

begin  quote
On Mon, 12 Apr 2004 22:17:30 +0100
Stuart Herbert <stuart@gentoo.org> wrote:

> On Monday 12 April 2004 21:58, Marius Mauch wrote:
> > On 04/12/04  Stuart Herbert wrote:
> > > Will Portage function without /var/cache/edb?
> >
> > Yes, but it will be _very_ slow.
> >
> > Marius
> 
> Then we shouldn't be putting this in /var/cache.  It's for temporary
> files,  and can be emptied at any time without warning.  This should
> really go  into /var/db.


even worse. the "world" file, and the configs are also stored in
/var/cache.  "not good".


//Spider



 


-- 
begin  .signature
Tortured users / Laughing in pain
See Microsoft KB Article Q265230 for more information.
end

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Drake Wyrm]

[-- Attachment #1: Type: text/plain, Size: 1145 bytes --]

On Mon, 2004-04-12, 23:26:06 +0200, in <20040412232606.51a010f7.spider@gentoo.org>, Spider <spider@gentoo.org> wrote:
> begin  quote
> On Mon, 12 Apr 2004 22:17:30 +0100
> Stuart Herbert <stuart@gentoo.org> wrote:
> 
> > On Monday 12 April 2004 21:58, Marius Mauch wrote:
> > > On 04/12/04  Stuart Herbert wrote:
> > > > Will Portage function without /var/cache/edb?
> > >
> > > Yes, but it will be _very_ slow.
> > >
> > > Marius
> > 
> > Then we shouldn't be putting this in /var/cache.  It's for temporary
> > files,  and can be emptied at any time without warning.  This should
> > really go  into /var/db.
> 
> 
> even worse. the "world" file, and the configs are also stored in
> /var/cache.  "not good".

If I understand FHS correctly, the "cache" stuff should go somewhere
under /var/tmp. /var/cache/edb/{config,counter,mtimedb,virtuals,world}
and /var/db/pkg should go somewhere under /var/lib. FHS specifically
states that /var/db shouldn't even exist.

-- 
Batou: Hey, Major... You ever hear of "human rights"?
Kusanagi: I understand the concept, but I've never seen it in action.
  --Ghost in the Shell

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 859 bytes --]

On 04/12/04  Stuart Herbert wrote:

> On Monday 12 April 2004 21:58, Marius Mauch wrote:
> > On 04/12/04  Stuart Herbert wrote:
> > > Will Portage function without /var/cache/edb?
> >
> > Yes, but it will be _very_ slow.
> >
> > Marius
> 
> Then we shouldn't be putting this in /var/cache.  It's for temporary
> files, and can be emptied at any time without warning.  This should
> really go into /var/db.

I don't understand you here: The cache files are temporary, they can be
removed at any time and will be regenerated when portage accesses the
ebuild, just the first run will be very slow (I could give you numbers
if I wouldn't run a world update atm).

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Stuart Herbert]

[-- Attachment #1: Type: text/plain, Size: 917 bytes --]

On Monday 12 April 2004 22:26, Marius Mauch wrote:
> I don't understand you here: The cache files are temporary, they can be
> removed at any time and will be regenerated when portage accesses the
> ebuild, just the first run will be very slow (I could give you numbers
> if I wouldn't run a world update atm).
>
> Marius

If you did this:

	rm -rf /var/cache/*

on your machine right now, portage would continue to work 100%?  There would 
be no loss of functionality at all?  Just the performance loss?

Best regards,
Stu
-- 
Stuart Herbert                                              stuart@gentoo.org
Gentoo Developer                                       http://www.gentoo.org/
Missed the php|cruise?             http://dev.gentoo.org/~stuart/cruise-2004/

GnuGP key id# F9AFC57C available from http://pgp.mit.edu
Key fingerprint = 31FB 50D4 1F88 E227 F319  C549 0C2F 80BA F9AF C57C
--

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Andrew Gaffney]

Stuart Herbert wrote:
> On Monday 12 April 2004 22:26, Marius Mauch wrote:
> 
>>I don't understand you here: The cache files are temporary, they can be
>>removed at any time and will be regenerated when portage accesses the
>>ebuild, just the first run will be very slow (I could give you numbers
>>if I wouldn't run a world update atm).
>>
>>Marius
> 
> 
> If you did this:
> 
> 	rm -rf /var/cache/*
> 
> on your machine right now, portage would continue to work 100%?  There would 
> be no loss of functionality at all?  Just the performance loss?

Well, if you did that, you'd kill your world file, virtuals, etc. You'd want to 'rm -rf 
/var/cache/edb/dep'.

-- 
Andrew Gaffney
Network Administrator
Skyline Aeronautics, LLC.
636-357-1548


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Stuart Herbert]

[-- Attachment #1: Type: text/plain, Size: 673 bytes --]

On Monday 12 April 2004 23:18, Andrew Gaffney wrote:
> Well, if you did that, you'd kill your world file, virtuals, etc. You'd
> want to 'rm -rf /var/cache/edb/dep'.

Then, again, this belongs in /var/db, and not /var/cache.  /var/cache is for 
*temporary* files only.

Best regards,
Stu
-- 
Stuart Herbert                                              stuart@gentoo.org
Gentoo Developer                                       http://www.gentoo.org/
Missed the php|cruise?             http://dev.gentoo.org/~stuart/cruise-2004/

GnuGP key id# F9AFC57C available from http://pgp.mit.edu
Key fingerprint = 31FB 50D4 1F88 E227 F319  C549 0C2F 80BA F9AF C57C
--

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 1005 bytes --]

On 04/12/04  Stuart Herbert wrote:

> On Monday 12 April 2004 22:26, Marius Mauch wrote:
> > I don't understand you here: The cache files are temporary, they can
> > be removed at any time and will be regenerated when portage accesses
> > the ebuild, just the first run will be very slow (I could give you
> > numbers if I wouldn't run a world update atm).
> >
> > Marius
> 
> If you did this:
> 
> 	rm -rf /var/cache/*
> 
> on your machine right now, portage would continue to work 100%?  There
> would be no loss of functionality at all?  Just the performance loss?

Ah, I misread your original reply. The cache files (which was the topic)
can stay there, I agree that the other files (world, virtuals) should be
moved.
Oh, portage should continue to work, although differently than you would
expect ;)

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 1234 bytes --]

On 04/12/04  Marius Mauch wrote:

> On 04/12/04  Stuart Herbert wrote:
> 
> > On Monday 12 April 2004 21:58, Marius Mauch wrote:
> > > On 04/12/04  Stuart Herbert wrote:
> > > > Will Portage function without /var/cache/edb?
> > >
> > > Yes, but it will be _very_ slow.
> > >
> > > Marius
> > 
> > Then we shouldn't be putting this in /var/cache.  It's for temporary
> > files, and can be emptied at any time without warning.  This should
> > really go into /var/db.
> 
> I don't understand you here: The cache files are temporary, they can
> be removed at any time and will be regenerated when portage accesses
> the ebuild, just the first run will be very slow (I could give you
> numbers if I wouldn't run a world update atm).

For the records:
# rm -rf /var/cache/edb/dep
# time emerge -epv system

real	0m57.258s
user	0m36.897s
sys	0m8.694s

# time emerge -epv system

real	0m8.785s
user	0m6.244s
sys	0m0.733s

That's on a Athlon XP 2600+ with no other resource-intensive processes
running. The slower your box the worse it gets.

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Re: [gentoo-dev] Idea for the portage maintainers

[brettholcomb]

One very nice thing about NOT being compressed is that I can work with the files as they are and not have to untar/zip/whatever first.  I can just look at the ebuilds, work with them, etc.  
> 
> From: Tom St Denis <tom@securescience.net>
> Date: 2004/04/12 Mon PM 12:36:15 GMT
> To: gentoo-dev@lists.gentoo.org
> Subject: Re: [gentoo-dev] Idea for the portage maintainers
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On April 12, 2004 08:23 am, Georgi Georgiev wrote:
> > maillog: 12/04/2004-08:03:13(-0400): Tom St Denis types
> >
> Also the portage files are not compressed.  If you take an entire directory 
> [again say app-text] and compress it you save bandwidth.  A quick check of 
> app-text on my box [updated last night] gives a 850K zip file [290K tar.bz2] 
> compared to 2.7M of raw data.  Obviously sending the entire zip would be 
> wasting more bandwidth when only small changes occur.
 


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On April 12, 2004 08:46 am, brettholcomb@charter.net wrote:
> One very nice thing about NOT being compressed is that I can work with the
> files as they are and not have to untar/zip/whatever first.  I can just
> look at the ebuilds, work with them, etc.

You could do that with a tool though...

emerge list app-text/tetex

Or even make a small "shell" like application so you could navigate the tree.

Anyways yeah going way off.  I think the best suggestion so far is to just 
loopback with a fs that uses small inodes.  While I probably could figure out 
how to do it, it probably wouldn't be a nice thing for newer users.

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAepLCsP+tEsHHY0ARAr/iAJ4kX/TfhKki3Ddup1A887+qYxKrkwCfScmr
1webOHm7Y8DvGgEsgX1q2Fs=
=CCx4
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 1104 bytes --]

On 04/12/04  Tom St Denis wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On April 12, 2004 08:46 am, brettholcomb@charter.net wrote:
> > One very nice thing about NOT being compressed is that I can work
> > with the files as they are and not have to untar/zip/whatever first.
> >  I can just
> > look at the ebuilds, work with them, etc.
> 
> You could do that with a tool though...
> 
> emerge list app-text/tetex
> 
> Or even make a small "shell" like application so you could navigate
> the tree.
> 
> Anyways yeah going way off.  I think the best suggestion so far is to
> just loopback with a fs that uses small inodes.  While I probably
> could figure out how to do it, it probably wouldn't be a nice thing
> for newer users.

Well, I still have the vision of modular portdb backends including the
current format, sql database and portagefs (which would solve all these
problems ;)

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I think a cool function [which I didn't see in the latest portage release] is 
a "snapshot" and restore set of functionality.  e.g. you can snapshot the 
current install set and later restore (by adding/removing packages) as 
required. 

I know I could have used this functionality before.  Like when I tried out 
GNOME and decided on KDE.... I still have GNOME code lying around.  There's 
probably a dozen other dependencies lying around from packages I tried out...

Also any plans to optimize the portage files?  80k small files amounts to huge 
waste of space.  

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAeTIXsP+tEsHHY0ARAmLJAJsFGqB8vIEipmMdQC00DEmvp4iBbACeMQm/
oTL7WjjZbYWWMaUTiwz8WiE=
=b3Fw
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Alexander Gretencord]

On Sunday 11 April 2004 13:55, Tom St Denis wrote:
> I think a cool function [which I didn't see in the latest portage release]
> is a "snapshot" and restore set of functionality.  e.g. you can snapshot
> the current install set and later restore (by adding/removing packages) as
> required.

What exactly do you mean by that? What irritates me is that "by 
adding/removing packages". If you just want a list of all installed packages 
you can already get it. What I'd understand by snapshotting is really 
packaging a set of packages and their deps. Kind of a '-b' but after merging.

> Also any plans to optimize the portage files?  80k small files amounts to
> huge waste of space.

Depends on your filesystem. With certain filesystems you could adjust the 
blocksize for /usr/portage/ to a smaller value (with a larger value 
for /usr/portage/distfiles of course) or use reiserfs with tail packing on.



Alex

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On April 12, 2004 06:45 am, Alexander Gretencord wrote:
> On Sunday 11 April 2004 13:55, Tom St Denis wrote:
> > I think a cool function [which I didn't see in the latest portage
> > release] is a "snapshot" and restore set of functionality.  e.g. you can
> > snapshot the current install set and later restore (by adding/removing
> > packages) as required.
>
> What exactly do you mean by that? What irritates me is that "by
> adding/removing packages". If you just want a list of all installed
> packages you can already get it. What I'd understand by snapshotting is
> really packaging a set of packages and their deps. Kind of a '-b' but after
> merging.

Well Ideally a command that simply emits a list of installed packages is what 
I'm talking about.  But specifically, it emits the list [and say GZIPs it at 
the same time] into a db of "restore points".  

So as another poster said I could do say 

emerge restore "11/04/04"

If I installed something today that I didn't like.  The trick is to make this 
painless for the user so the user won't scream in horror and go install 
Windows.

> > Also any plans to optimize the portage files?  80k small files amounts to
> > huge waste of space.
>
> Depends on your filesystem. With certain filesystems you could adjust the
> blocksize for /usr/portage/ to a smaller value (with a larger value
> for /usr/portage/distfiles of course) or use reiserfs with tail packing on.

Oh, ok so I'll just format my disk, reinstall Gentoo from scratch so that I 
can not waste 200M of space on 80k small files.

That's not really user friendly.  Could have done a JAR like setup for each 
dir of the tree.  e.g. all of app-text be one huge ZIP file [with no 
compression].  Such a setup might be a little slower to add/remove files but 
would waste less space.

The idea would make a little sense though in practice.  When I do "emerge 
sync" instead of fetching 1000s of small files I just check the timestamp on 
the directory zips and download them wholesale.  [Ok so maybe compression 
makes sense here].

That way as a user I don't have to worry about inserting/deleting files from a 
zip [which would take a while] only the server has todo it.  In fact the 
server could still use the "many files" approach and just zip on the fly when 
a user syncs.

I know that would make things a bit more complicated since mirrors update via 
"sync" as well.  I guess you could just have two types of portage trees.  
e.g. "packaged" and "straight" or something like that.  End users would use 
the "packaged" type [e.g. zip per category] and mirrors would use straight 
[e.g. all 80k files].

Anyways, I'm just throwing out ideas here.  Using the zip approach makes sense 
for end users.  First, it makes for faster syncing [less smaller files means 
fewer metadata commands] and wastes much less space [and would be faster 
since the kernel could cache it quicker!].

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAeoWDsP+tEsHHY0ARAg/dAJwMV7cUdGl8GJ/qA0StH9RFwRajCACfWqL4
hv++FVTFxDrzXKCKrrN1s4Y=
=wEEo
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Georgi Georgiev]

[-- Attachment #1: Type: text/plain, Size: 1203 bytes --]

maillog: 12/04/2004-08:03:13(-0400): Tom St Denis types
> Oh, ok so I'll just format my disk, reinstall Gentoo from scratch so that I 
> can not waste 200M of space on 80k small files.

You can always use a loopback device with a filesystem of your choice.
http://forums.gentoo.org/viewtopic.php?t=68215

> That's not really user friendly.  Could have done a JAR like setup for each 
> dir of the tree.  e.g. all of app-text be one huge ZIP file [with no 
> compression].  Such a setup might be a little slower to add/remove files but 
> would waste less space.
> 
> The idea would make a little sense though in practice.  When I do "emerge 
> sync" instead of fetching 1000s of small files I just check the timestamp on 
> the directory zips and download them wholesale.  [Ok so maybe compression 
> makes sense here].

How is this method faster or in any way better than rsync?

-- 
(    Georgi Georgiev   (   I don't need to compromise my principles,   (
 )    chutz@gg3.net     ) because they don't have the slightest         )
(   +81(90)6266-1163   (  bearing on what happens to me anyway. --     (
 ) -------------------  ) Calvin                                        )

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On April 12, 2004 08:23 am, Georgi Georgiev wrote:
> maillog: 12/04/2004-08:03:13(-0400): Tom St Denis types
>
> > Oh, ok so I'll just format my disk, reinstall Gentoo from scratch so that
> > I can not waste 200M of space on 80k small files.
>
> You can always use a loopback device with a filesystem of your choice.
> http://forums.gentoo.org/viewtopic.php?t=68215

That's a potential solution but a bit out of the way for the user don't you 
think?  Isn't the point of quality software to attract users simply by having 
merit?

> > That's not really user friendly.  Could have done a JAR like setup for
> > each dir of the tree.  e.g. all of app-text be one huge ZIP file [with no
> > compression].  Such a setup might be a little slower to add/remove files
> > but would waste less space.
> >
> > The idea would make a little sense though in practice.  When I do "emerge
> > sync" instead of fetching 1000s of small files I just check the timestamp
> > on the directory zips and download them wholesale.  [Ok so maybe
> > compression makes sense here].
>
> How is this method faster or in any way better than rsync?

Less metadata?  If I want to download 1000 files off your site I have to say 

+GET /file1
- -HTTP/1.1 200 OK
... data
+GET /file2
- -HTTP/1.1 200 OK
... data
+GET /file3
- -HTTP/1.1 200 OK
... data
ETC

The overhead translates on both sides too.  E.g. My FS now has to find room 
for and enter in 1000s of files.  The server side has to locate 1000s of 
files.

Also the portage files are not compressed.  If you take an entire directory 
[again say app-text] and compress it you save bandwidth.  A quick check of 
app-text on my box [updated last night] gives a 850K zip file [290K tar.bz2] 
compared to 2.7M of raw data.  Obviously sending the entire zip would be 
wasting more bandwidth when only small changes occur.

Specifically my point isn't to use zip but to find a way to cluster files 
[automatically].  A loopback on the client side would work but it shouldn't 
be manual.

Anyways I'm not trying to rock the boat here.  portage for the most part does 
work.  It's just not very scaleable [I recall a sync not taking so f'ing long 
to update the cache, or a "world" update being lightning fast...]

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD4DBQFAeo0/sP+tEsHHY0ARAgGYAJ9g6vrv4rwB7vkxKTiuz0fpKTiaXwCXYeWL
PSsib2TAPxz1hDwq1G6L6A==
=IUg+
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[N. Owen Gunden]

On Mon, Apr 12, 2004 at 08:36:15AM -0400, Tom St Denis wrote:
> > How is this method faster or in any way better than rsync?
> 
> Less metadata?  If I want to download 1000 files off your site I have to say 
> 
> +GET /file1
> - -HTTP/1.1 200 OK
> ... data
> +GET /file2
> - -HTTP/1.1 200 OK
> ... data
> +GET /file3
> - -HTTP/1.1 200 OK
> ... data
> ETC

Hmm... it looks like you need to do a little research into how rsync
transfers files.  It's much more efficient than what you're thinking.
And it certainly doesn't use HTTP.

 - O

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Troy Dack]

On Mon, 2004-04-12 at 22:36, Tom St Denis wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On April 12, 2004 08:23 am, Georgi Georgiev wrote:
> > maillog: 12/04/2004-08:03:13(-0400): Tom St Denis types
> >
> > > Oh, ok so I'll just format my disk, reinstall Gentoo from scratch so that
> > > I can not waste 200M of space on 80k small files.
> >
> > You can always use a loopback device with a filesystem of your choice.
> > http://forums.gentoo.org/viewtopic.php?t=68215
> 
> That's a potential solution but a bit out of the way for the user don't you 
> think?  Isn't the point of quality software to attract users simply by having 
> merit?
> 
> > > That's not really user friendly.  Could have done a JAR like setup for
> > > each dir of the tree.  e.g. all of app-text be one huge ZIP file [with no
> > > compression].  Such a setup might be a little slower to add/remove files
> > > but would waste less space.
> > >
> > > The idea would make a little sense though in practice.  When I do "emerge
> > > sync" instead of fetching 1000s of small files I just check the timestamp
> > > on the directory zips and download them wholesale.  [Ok so maybe
> > > compression makes sense here].
> >
> > How is this method faster or in any way better than rsync?
> 
> Less metadata?  If I want to download 1000 files off your site I have to say 
> 
> +GET /file1
> - -HTTP/1.1 200 OK
> ... data
> +GET /file2
> - -HTTP/1.1 200 OK
> ... data
> +GET /file3
> - -HTTP/1.1 200 OK
> ... data
> ETC
>
> The overhead translates on both sides too.  E.g. My FS now has to find room 
> for and enter in 1000s of files.  The server side has to locate 1000s of 
> files.
> 
> Also the portage files are not compressed.  If you take an entire directory 
> [again say app-text] and compress it you save bandwidth.  A quick check of 
> app-text on my box [updated last night] gives a 850K zip file [290K tar.bz2] 
> compared to 2.7M of raw data.  Obviously sending the entire zip would be 
> wasting more bandwidth when only small changes occur.

Rsync compresses the information it sends, additionally (and most
importantly) rsync only sends the differences, not the whole file, so it
is even more efficient. Visit rsync.samba.org and have a read about how
it all works.

Another point against a monolithic zip containing all the ebuilds (or
even per directory zips) is the performance hit that slow machines would
take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
PII-400 in the corner)

-- 
Troy Dack					http://linux.tkdack.com
<troy@tkdack.com>				http://webportage.sf.net

Public Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x4D90BE3C
Key fingerprint = 1F3D 6C15 16AA 09D5 0C96  92E5 FD89 16F9 4D90 BE3C


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Jason Stubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 13 April 2004 00:12, Troy Dack wrote:
> Rsync compresses the information it sends, additionally (and most
> importantly) rsync only sends the differences, not the whole file, so it
> is even more efficient. Visit rsync.samba.org and have a read about how
> it all works.

Rsync does only send differences by default but emerge disables this behaviour 
simply because most of the files are about 1kb in size. The overhead CPU on 
both client and server is not worth the relatively small gain of doing it.

Regards,
Jason Stubbs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iQCVAwUBQHqyo1oikN4/5jfsAQJxIAP+Mjv7Auw0r0QOU8ov7EHJ/qlmh8JqKc2Z
/Bd8+hLsuGkmtrL38WRwtDiPKXashbPF+VYQnHSEO8+2n8Ewivpp84iefgJGTRzX
SP5m50GunidHQ96ulS6Sffaemws0iEVy16ovnw6vesccxZ2uQcJpHtTNDHeW+fZx
wOvNGt7lBik=
=x1zJ
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Andrew Gaffney]

Troy Dack wrote:
> Another point against a monolithic zip containing all the ebuilds (or
> even per directory zips) is the performance hit that slow machines would
> take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
> PII-400 in the corner)

Or my little P233 Thinkpad...

-- 
Andrew Gaffney
Network Administrator
Skyline Aeronautics, LLC.
636-357-1548


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Todd Berman]

On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
> Troy Dack wrote:
> > Another point against a monolithic zip containing all the ebuilds (or
> > even per directory zips) is the performance hit that slow machines would
> > take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
> > PII-400 in the corner)
> 
> Or my little P233 Thinkpad...
> 

And with the current setup of writing thousands of 1K files that little
p233 thinkpad really flys i bet...

--Todd


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Andrew Gaffney]

Todd Berman wrote:
> On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
> 
>>Troy Dack wrote:
>>
>>>Another point against a monolithic zip containing all the ebuilds (or
>>>even per directory zips) is the performance hit that slow machines would
>>>take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
>>>PII-400 in the corner)
>>
>>Or my little P233 Thinkpad...
>>
> 
> 
> And with the current setup of writing thousands of 1K files that little
> p233 thinkpad really flys i bet...

I'm not sure if that was supposed to be sarcastic, but yes, it does fly. It only takes 
slightly longer to sync than my Athlon 1.3GHz desktop. The only part that takes forever is 
updating the portage cache. That's why I just use a NFS shared portage tree from my 
desktop machine now.

-- 
Andrew Gaffney
Network Administrator
Skyline Aeronautics, LLC.
636-357-1548


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Todd Berman]

On Mon, 2004-12-04 at 11:59 -0500, Andrew Gaffney wrote:
> Todd Berman wrote:
> > On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
> > 
> >>Troy Dack wrote:
> >>
> >>>Another point against a monolithic zip containing all the ebuilds (or
> >>>even per directory zips) is the performance hit that slow machines would
> >>>take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
> >>>PII-400 in the corner)
> >>
> >>Or my little P233 Thinkpad...
> >>
> > 
> > 
> > And with the current setup of writing thousands of 1K files that little
> > p233 thinkpad really flys i bet...
> 
> I'm not sure if that was supposed to be sarcastic, but yes, it does fly. It only takes 
> slightly longer to sync than my Athlon 1.3GHz desktop. The only part that takes forever is 
> updating the portage cache. That's why I just use a NFS shared portage tree from my 
> desktop machine now.

In a way it was and in a way it wasn't. I honestly don't understand how
you can explain to me that a compression-less zip file be any slower
than the current setup.

With compression I could understand, but without I don't think any speed
difference would be noticeable. However, even with compression the
operation should be fairly fast. zip is not like a tar.gz or tar.bz2,
ie, you can read a file out of it and search through it fairly fast, and
you don't have to uncompress the entire archive to get a single file.

--Todd


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Andrew Gaffney]

Todd Berman wrote:
> On Mon, 2004-12-04 at 11:59 -0500, Andrew Gaffney wrote:
> 
>>Todd Berman wrote:
>>
>>>On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
>>>
>>>
>>>>Troy Dack wrote:
>>>>
>>>>
>>>>>Another point against a monolithic zip containing all the ebuilds (or
>>>>>even per directory zips) is the performance hit that slow machines would
>>>>>take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
>>>>>PII-400 in the corner)
>>>>
>>>>Or my little P233 Thinkpad...
>>>>
>>>
>>>
>>>And with the current setup of writing thousands of 1K files that little
>>>p233 thinkpad really flys i bet...
>>
>>I'm not sure if that was supposed to be sarcastic, but yes, it does fly. It only takes 
>>slightly longer to sync than my Athlon 1.3GHz desktop. The only part that takes forever is 
>>updating the portage cache. That's why I just use a NFS shared portage tree from my 
>>desktop machine now.
> 
> In a way it was and in a way it wasn't. I honestly don't understand how
> you can explain to me that a compression-less zip file be any slower
> than the current setup.
> 
> With compression I could understand, but without I don't think any speed
> difference would be noticeable. However, even with compression the
> operation should be fairly fast. zip is not like a tar.gz or tar.bz2,
> ie, you can read a file out of it and search through it fairly fast, and
> you don't have to uncompress the entire archive to get a single file.

Even without compression, there is still a little bit of overhead with having ebuilds and 
such contained in ZIP files. The overhead comes into play both during sync and emerge. It 
wouldn't be noticable on a fast machine (e.g. my Athlon 1.3GHz desktop) but it would make 
an operation like 'emerge -uDpv world' take quite a bit longer on my Thinkpad.

As someone else pointed out, the ZIP files wouldn't make syncs faster either. Rsync 
transfers only the differences in files and not the entire app-text directory or whatever 
category you're working with.

As another person pointed out, it would be more difficult to modify ebuilds in the tree by 
hand when they're contained in ZIP files.

It was a good idea, but I don't think it is practical.

-- 
Andrew Gaffney
Network Administrator
Skyline Aeronautics, LLC.
636-357-1548


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Todd Berman]

On Mon, 2004-12-04 at 12:17 -0500, Andrew Gaffney wrote:
> Todd Berman wrote:
> > On Mon, 2004-12-04 at 11:59 -0500, Andrew Gaffney wrote:
> > 
> >>Todd Berman wrote:
> >>
> >>>On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
> >>>
> >>>
> >>>>Troy Dack wrote:
> >>>>
> >>>>
> >>>>>Another point against a monolithic zip containing all the ebuilds (or
> >>>>>even per directory zips) is the performance hit that slow machines would
> >>>>>take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
> >>>>>PII-400 in the corner)
> >>>>
> >>>>Or my little P233 Thinkpad...
> >>>>
> >>>
> >>>
> >>>And with the current setup of writing thousands of 1K files that little
> >>>p233 thinkpad really flys i bet...
> >>
> >>I'm not sure if that was supposed to be sarcastic, but yes, it does fly. It only takes 
> >>slightly longer to sync than my Athlon 1.3GHz desktop. The only part that takes forever is 
> >>updating the portage cache. That's why I just use a NFS shared portage tree from my 
> >>desktop machine now.
> > 
> > In a way it was and in a way it wasn't. I honestly don't understand how
> > you can explain to me that a compression-less zip file be any slower
> > than the current setup.
> > 
> > With compression I could understand, but without I don't think any speed
> > difference would be noticeable. However, even with compression the
> > operation should be fairly fast. zip is not like a tar.gz or tar.bz2,
> > ie, you can read a file out of it and search through it fairly fast, and
> > you don't have to uncompress the entire archive to get a single file.
> 
> Even without compression, there is still a little bit of overhead with having ebuilds and 
> such contained in ZIP files. The overhead comes into play both during sync and emerge. It 
> wouldn't be noticable on a fast machine (e.g. my Athlon 1.3GHz desktop) but it would make 
> an operation like 'emerge -uDpv world' take quite a bit longer on my Thinkpad.
> 

This is going to be a slow operation on that thinkpad regardless.

> As someone else pointed out, the ZIP files wouldn't make syncs faster either. Rsync 
> transfers only the differences in files and not the entire app-text directory or whatever 
> category you're working with.
> 

Portage does not use this as it takes more cpu server and client side to
figure out what changed with 80 thousand 1K files than it does to just
transfer the files across.

> As another person pointed out, it would be more difficult to modify ebuilds in the tree by 
> hand when they're contained in ZIP files.
> 

This could easily be dealt with. the cvs->rsync process could create zip
files, and portage would use the zip stuff by default for /usr/portage/
and use the old directory structure code for your overlays.

> It was a good idea, but I don't think it is practical.

It is absolutely practical, and I think would cut down on the time taken
to sync (because now you could turn compression back on, and actually
transfer just changes because you wouldnt be syncing 80 thousand files).
It might potentially have small speed loses when running emerge -vuDp
world, but as that process takes a long time regardless, I don't think
it would be noticeable. The time it takes to read a single file out of
an uncompressed zip is not noticeably longer than the time it takes to
read a single file off the hard drive.

Note, as of writing this the portage tree takes up 333MB of space on my
hard drive and a compression-less zipfile of the tree takes up 74MB. To
me, the potential savings are incredible.

--Todd


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Jason Stubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 13 April 2004 02:39, Todd Berman wrote:
> I think would cut down on the time taken to sync (because now you could turn 
> compression back on, and actually transfer just changes because you wouldnt 
> be syncing 80 thousand files).  

If this is referring to my other comment, I never said that compression is 
disabled. It is not and is definately a real benefit. I find that the data is 
compressed at approximately 5:1 on average.

Regards,
Jason Stubbs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iQCVAwUBQHs8rFoikN4/5jfsAQJ5zQP/dpSBTGEBycUp6/dwLLO3OHwhkSLrNHAs
yoTcy2u4ZEb9yPTaqc0iwMd00PYYFuBGnsbDmScke1gVA7qewUWCBLTYae7fk7Zm
79Lgk8n4ZAWmw5sVYJVfkzT81o7m56pTv2G4X45eYf8wvRQbTtsY6/1WCcpaaDSC
V8N/Npw8uy0=
=t2gz
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Todd Berman]

On Tue, 2004-13-04 at 10:04 +0900, Jason Stubbs wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On Tuesday 13 April 2004 02:39, Todd Berman wrote:
> > I think would cut down on the time taken to sync (because now you could turn 
> > compression back on, and actually transfer just changes because you wouldnt 
> > be syncing 80 thousand files).  
> 
> If this is referring to my other comment, I never said that compression is 
> disabled. It is not and is definately a real benefit. I find that the data is 
> compressed at approximately 5:1 on average.
> 
> Regards,
> Jason Stubbs

Yeah, sorry, differences is what i mean.

--Todd

> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> 
> iQCVAwUBQHs8rFoikN4/5jfsAQJ5zQP/dpSBTGEBycUp6/dwLLO3OHwhkSLrNHAs
> yoTcy2u4ZEb9yPTaqc0iwMd00PYYFuBGnsbDmScke1gVA7qewUWCBLTYae7fk7Zm
> 79Lgk8n4ZAWmw5sVYJVfkzT81o7m56pTv2G4X45eYf8wvRQbTtsY6/1WCcpaaDSC
> V8N/Npw8uy0=
> =t2gz
> -----END PGP SIGNATURE-----
> 
> --
> gentoo-dev@gentoo.org mailing list
> 


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On April 12, 2004 12:59 pm, Andrew Gaffney wrote:
> Todd Berman wrote:
> > On Mon, 2004-12-04 at 11:22 -0500, Andrew Gaffney wrote:
> >>Troy Dack wrote:
> >>>Another point against a monolithic zip containing all the ebuilds (or
> >>>even per directory zips) is the performance hit that slow machines would
> >>>take, not everybody runs gentoo on a 2GHz plus machine (eg: my little
> >>>PII-400 in the corner)
> >>
> >>Or my little P233 Thinkpad...
> >
> > And with the current setup of writing thousands of 1K files that little
> > p233 thinkpad really flys i bet...
>
> I'm not sure if that was supposed to be sarcastic, but yes, it does fly. It
> only takes slightly longer to sync than my Athlon 1.3GHz desktop. The only
> part that takes forever is updating the portage cache. That's why I just
> use a NFS shared portage tree from my desktop machine now.

That isn't a solution though.  

"Gentoo, the distributed home computer operating system.  "

"Gentoo, takes 2379MIPS to sync within a week."

etc...

I mean if it's slow now what happens in [say] 2008 when portage has 160k files 
or something... I don't know exactly how portage works but from when I 
started using gentoo [~60k files in portage] to now [~80k files in portage] 
it's definitely not just a linear amount slower.

While I'm at it what of security?  I mean how do I know the files on the 
mirror I get them from are from the CVS?

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAes1FsP+tEsHHY0ARApG8AKCQhFU2mkTIHo8LIB8pkZzINmRQaACcDuC9
c8XWfm8lMvKVXobONXSoWLE=
=D9Cy
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Norberto Bensa]

Tom St Denis wrote:
> I started using gentoo [~60k files in portage] to now [~80k files in
> portage] it's definitely not just a linear amount slower.

rsync isn't slow. "emerge --sync" is slow in the portage cache updates stage 
(I really want to know what portage cache update does :-/ )

Regards,
Norberto

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Tom St Denis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On April 12, 2004 01:19 pm, Norberto Bensa wrote:
> Tom St Denis wrote:
> > I started using gentoo [~60k files in portage] to now [~80k files in
> > portage] it's definitely not just a linear amount slower.
>
> rsync isn't slow. "emerge --sync" is slow in the portage cache updates
> stage (I really want to know what portage cache update does :-/ )

Syncs can be slow too.  Sometimes they zoom but sometimes it will say things 
like

200 files ......

and twirl..... and twirl......and twirl......

Point is portage is useable and works decent.  Updating is a bitch and 
occasionally a "world" update takes way too long [to get the package list].  
It's a manner of being scaleable.  Portage isn't.

Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAetA0sP+tEsHHY0ARAuJIAJ9jfAozHPtz2o49wiNQdni1FYU5FgCZAefS
+qb63BEv19W73J6HN6v06WA=
=CRvw
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Chris Bainbridge]

On Monday 12 April 2004 13:03, Tom St Denis wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Oh, ok so I'll just format my disk, reinstall Gentoo from scratch so that I
> can not waste 200M of space on 80k small files.
>
> That's not really user friendly.  Could have done a JAR like setup for each
> dir of the tree.  e.g. all of app-text be one huge ZIP file [with no
> compression].  Such a setup might be a little slower to add/remove files
> but would waste less space.
>
> The idea would make a little sense though in practice.  When I do "emerge
> sync" instead of fetching 1000s of small files I just check the timestamp
> on the directory zips and download them wholesale.  [Ok so maybe
> compression makes sense here].

I once posted a test of using rsync with a single large file versus lots of 
small ones on the gentoo forums 
(http://forums.gentoo.org/viewtopic.php?t=10108). I was more concerned with 
bandwidth than speed, but I guess they're connected. Using one big file gave 
a 35% improvement over lots of small files. The problem then was how to 
generate one big file for syncing. In the end I got bored and decided it was 
rsyncs problem for not being as efficient with the small files ;-)

The test of rsyncing uncompressed distfile sources rather than downloading 
them gave an 88% improvement in bandwidth. Problem here is that we would need 
lots of disk space for the uncompressed files.. so its a tradeoff, if your 
bandwidth is limited or expensive and computation and disk space are cheap  
(eg. dialup), then it seems like a good idea.

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Chris Bainbridge]

On Tuesday 13 April 2004 13:18, Chris Bainbridge wrote:
> The test of rsyncing uncompressed distfile sources rather than downloading
> them gave an 88% improvement in bandwidth.

I should add that rsyncing from an old distfiles tarball to the new one used 
323k downstream + 58k up, when downloading the new bzip2 file would've used  
about 2.6MB downstream. So uncompressed tar+rsync is over 8x more bandwidth 
efficient.  

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Idea for the portage maintainers

[Senor Rodgman]

Hi,

I implemented this some time ago. Have a look at bug #40127
(http://bugs.gentoo.org/show_bug.cgi?id=40127). It lets you do stuff like:

emerge --restore 10/03/2003-18:00

Unfortunately what with the portage rewrite it doesn't look like this will
get merged into mainstream portage. Have a look, it might be what you
want.

dave

On Sun, 11 Apr 2004, Tom St Denis wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> I think a cool function [which I didn't see in the latest portage release] is 
> a "snapshot" and restore set of functionality.  e.g. you can snapshot the 
> current install set and later restore (by adding/removing packages) as 
> required. 
> 
> I know I could have used this functionality before.  Like when I tried out 
> GNOME and decided on KDE.... I still have GNOME code lying around.  There's 
> probably a dozen other dependencies lying around from packages I tried out...
> 
> Also any plans to optimize the portage files?  80k small files amounts to huge 
> waste of space.  
> 
> Tom
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> 
> iD8DBQFAeTIXsP+tEsHHY0ARAmLJAJsFGqB8vIEipmMdQC00DEmvp4iBbACeMQm/
> oTL7WjjZbYWWMaUTiwz8WiE=
> =b3Fw
> -----END PGP SIGNATURE-----
> 
> --
> gentoo-dev@gentoo.org mailing list
> 
> 

-- 

--
gentoo-dev@gentoo.org mailing list