Delta-ISO: Building prozess howto?
Hello, for the SuSE-developer-versions there are Delta-ISO on the mirrors. I get them for the OpenSuSE 10.0 and felt, that this will be a good way for another CD-Project because it saves traffic. I hope You could point me to a howto. I need a documentation with more examples than the delta manpage. Regards, Andreas -- ## Content Developer OpenOffice.org: lang/DE ## Freie Office-Suite für Linux, Mac, Windows, Solaris ## http://de.openoffice.org ## Meine Seite http://www.amantke.de
Hello community ! Andreas Mantke wrote:
for the SuSE-developer-versions there are Delta-ISO on the mirrors. I get them for the OpenSuSE 10.0 and felt, that this will be a good way for another CD-Project because it saves traffic. I hope You could point me to a howto. I need a documentation with more examples than the delta manpage.
From my point of view it is a straight forward process. "makedeltaiso" is used to create a delta from two iso's (let's call these "source" and "target"; of course these should be of the same trace, e.g. CD1 of beta3 and beta4). "applydeltaiso" is used to cretae an iso (let's call it "target") from its "source" and an appropriate delta-iso (e.g. bulding beta4.iso from beta3.iso and beta3-beta4.delta.iso). To create a delta iso you simply need both iso's you want to build a delta iso from. If you are looking for a way to build a delta iso without having the second iso I fear there is no way for ... -- Never give up ! Best regards, Reinhard.
Hi, Reinhard Gimbel schrieb:
To create a delta iso you simply need both iso's you want to build a delta iso from.
And LOTS of RAM and CPU power. I'd recommend 4 GB RAM as minimum and the fastest processor you can get. It might still take a day or more if you're processing DVD isos.
If you are looking for a way to build a delta iso without having the second iso I fear there is no way for ...
Yes, because applydeltaiso is there to reconstruct a full iso. If you only want the files, you could try *deltarpm. Regards, Carl-Daniel -- http://www.hailfinger.org/
On Tue, 6 Dec 2005, Carl-Daniel Hailfinger wrote:
Reinhard Gimbel schrieb:
To create a delta iso you simply need both iso's you want to build a delta iso from.
And LOTS of RAM and CPU power. I'd recommend 4 GB RAM as minimum and the fastest processor you can get. It might still take a day or more if you're processing DVD isos.
Yeah + x86_64 would be highly recommended.
If you are looking for a way to build a delta iso without having the second iso I fear there is no way for ...
Yes, because applydeltaiso is there to reconstruct a full iso. If you only want the files, you could try *deltarpm.
Please note: The deltaiso stuff only works for rpm package on the ISOs - all the non-rpm stuff will get copied 1:1 from source to dest (if source is the original ISO and dest the deltaiso you are creating). Regards Christoph
Hello, first thanks for the information. Am Dienstag, 6. Dezember 2005 08:54 schrieb Christoph Thiel:
On Tue, 6 Dec 2005, Carl-Daniel Hailfinger wrote:
Reinhard Gimbel schrieb:
To create a delta iso you simply need both iso's you want to build a delta iso from.
And LOTS of RAM and CPU power. I'd recommend 4 GB RAM as minimum and the fastest processor you can get. It might still take a day or more if you're processing DVD isos.
My box have only 1 GB RAM. I don't know if this is enought to make a delta-cd-iso. What do you mean?
Yeah + x86_64 would be highly recommended.
There is a Athlon 64 3000+ inside the box.
If you are looking for a way to build a delta iso without having the second iso I fear there is no way for ...
Yes, because applydeltaiso is there to reconstruct a full iso. If you only want the files, you could try *deltarpm.
That's not the problem. I had first to build the new image. The old image is also on the box. So there are both ISO on the harddisk of the PC.
Please note: The deltaiso stuff only works for rpm package on the ISOs - all the non-rpm stuff will get copied 1:1 from source to dest (if source is the original ISO and dest the deltaiso you are creating).
This is very problematic. There are not only rpms on the cd. We put also binarys for linux, windows and other OS on the image. And there are textfiles (txt and html) on the cd. If I understood your words in the right way, delta only works with the rpm-stuff. The binarys and the text-files don't become smaler. Regards, Andreas -- ## Content Developer OpenOffice.org: lang/DE ## Freie Office-Suite für Linux, Mac, Windows, Solaris ## http://de.openoffice.org ## Meine Seite http://www.amantke.de
On Tue, Dec 06, 2005 at 06:57:34PM +0100, Andreas Mantke wrote:
My box have only 1 GB RAM. I don't know if this is enought to make a delta-cd-iso. What do you mean?
Depends on the CD data. Delta generation needs about three times the size of the data it is working on as memory. Just give it a try.
This is very problematic. There are not only rpms on the cd. We put also binarys for linux, windows and other OS on the image. And there are textfiles (txt and html) on the cd. If I understood your words in the right way, delta only works with the rpm-stuff. The binarys and the text-files don't become smaler.
No, that's wrong. The deltaiso algorithm also works with the other data on CD. It won't help you much if the data is compressed, though, as the delta between two compressed files (that are not identical) is quite big. As I said, just give it a try. Cheers, Michael. -- Michael Schroeder mls@suse.de main(_){while(_=~getchar())putchar(~_-1/(~(_|32)/13*2-11)*13);}
Hello, Am Dienstag, 6. Dezember 2005 19:04 schrieb Michael Schroeder:
On Tue, Dec 06, 2005 at 06:57:34PM +0100, Andreas Mantke wrote:
My box have only 1 GB RAM. I don't know if this is enought to make a delta-cd-iso. What do you mean?
Depends on the CD data. Delta generation needs about three times the size of the data it is working on as memory. Just give it a try.
Yes, but IIRC we are talking about the _biggest file_ on the CD, not the whole CD. (For RPMs, calculate with the _uncompressed_ size.) I guess you have a chance with 1 GB Ram. As Michael already said:
As I said, just give it a try.
;-) BTW: If your system keeps working for days, it's a clear sign of too less RAM. makedelta* from swap is terribly slow... Regards, Christian Boltz -- Postings sind nichts weiter als Kondensationskerne. Mit etwas Glück schlägt sich eine Diskussion an ihnen nieder die sich schon seit Tagen zusammengeballt hat. Oder aber wir haben trockenes Wetter und nix passiert. [Cornell Binder in dafu-l]
Hello, Am Dienstag, 6. Dezember 2005 22:02 schrieb Christian Boltz:
Hello,
Am Dienstag, 6. Dezember 2005 19:04 schrieb Michael Schroeder:
On Tue, Dec 06, 2005 at 06:57:34PM +0100, Andreas Mantke wrote:
My box have only 1 GB RAM. I don't know if this is enought to make a delta-cd-iso. What do you mean?
Depends on the CD data. Delta generation needs about three times the size of the data it is working on as memory. Just give it a try.
Yes, but IIRC we are talking about the _biggest file_ on the CD, not the whole CD. (For RPMs, calculate with the _uncompressed_ size.)
I guess you have a chance with 1 GB Ram. As Michael already said:
As I said, just give it a try.
I tried it and run into an error. makedelta reads the old and the new iso-image (each about 700 MB) and stops at delta-diffs. makedelta told me that there are not enough memory.
;-)
BTW: If your system keeps working for days, it's a clear sign of too less RAM. makedelta* from swap is terribly slow...
I don't had to wait long. The error displayed after 2 Minutes ;-) I think that I need about 3 MB memory for this action. Regards, Andreas -- ## Content Developer OpenOffice.org: lang/DE ## Freie Office-Suite für Linux, Mac, Windows, Solaris ## http://de.openoffice.org ## Meine Seite http://www.amantke.de
On Tue, Dec 06, 2005 at 12:48:40AM +0100, Carl-Daniel Hailfinger wrote:
Reinhard Gimbel schrieb:
To create a delta iso you simply need both iso's you want to build a delta iso from.
And LOTS of RAM and CPU power. I'd recommend 4 GB RAM as minimum and the fastest processor you can get. It might still take a day or more if you're processing DVD isos.
It shouldn't take a day for a DVD, I think 10 hours are a bit more realistic. Cheers, Michael. -- Michael Schroeder mls@suse.de main(_){while(_=~getchar())putchar(~_-1/(~(_|32)/13*2-11)*13);}
participants (6)
-
Andreas Mantke -
Carl-Daniel Hailfinger -
Christian Boltz -
Christoph Thiel -
Michael Schroeder -
Reinhard Gimbel