For LUG people in Delhi and beyond, who volunteer to copy and distribute Linux distributions, there is often a problem which to some degree affects this particular service.
Whenever a distribution ISO is offered for download, the ISO creators also include an MD5SUM checksum for people to verify both the authenticity and integrity of the image. This is a check against:
1. Somebody distributing malware in the name of the original CD
2. Problems in the download which can corrupt the CD image.
So after downloading the image, one can simply run the md5sum program on these images to ensure that they are getting an uncorrupted image to write on the CD.(Of course there still is the danger of the writing process itself introducing errors).
Generally a few people actually download the CD image, and most of the distribution in the LUG happens by exchanging/lending the written CDs. When a distribution volunteer first gets a copy of the CD, he/she uses his favorite program (dd/xcdroast/k3b) etc to copy the image to the hard disk. Later on, this image is used to write the copy for those who request for it.
Now the problem is that the whole process of copying of CDs from one media to another, introduces artifacts in the image which cause the CD image to start differing from the original downloadable image. The results show in wildly different md5sums of the first generation and second generation copies.
Mind you that such difference in MD5SUMs is no indication of the worthiness of the CD which can very well be as functional as the original.
However, difference in MD5SUMs can *also* be due the minuscule damages to the CD causing just a few files of a CD to be corrupt. These kind of damages can only be spotted when these particular files are accessed. Or during installation, some distribs offer a media check to ascertain the integrity of the CD.
So a volunteer CD distributor like me have no way to find out whether a difference in Md5sum is because the image is corrupted or because the copying process modified the image in some harmless way.
I personally faced this problem after one of my red hat 9 CD images came out with an integrity problem which only showed up if you run media check or try to install a small subset of the optional packages.
To solve this problem, I decided to create distribution CDs for distributions. This is nothing but a CD image containing a CD image.
Each of the CD isos are packages along with the MD5SUM into another CD image. As an example, the distrib CD in my CD drive is:
[EMAIL PROTECTED] redhat9]$ ls /mnt/cdrom MD5SUM shrike-i386-disc3.iso
This enables the CD distributor to copy the image to the hard disc and compare with the official MD5SUMs.
This seems to be the only foolproof way to assure those coming for CD copies, that their new CDs are really copies of the original ones, and not some damaged ones.
Yes, it seems like a terrible waste - the CDs are not bootable and the contents not directly usable. But again, not everybody needs to have these CDs - just people volunteering to write these CDs or organisations which use a number of copies of these distribs.
Probably if somebody has a DVD writer, they can actually write a DVD with the all the ISOs for the distribution(s) and their MD5SUM files inside it.
Every distributor need not burn a copy of this CD for himself - there can just be one copy from where every one can copy the images to their hard disks, and pass the CD on to others.
Thanks to the kind guys at Esys esp. Yashpal Nagar and his colleague, I managed to get my hands on an authentic set of RHL9 ISO images and proceeded to create distributions CDs. This enables both me and Yash to offer redhat 9 CD copies from "authentic" "official" RHL9 ISO images.
Those who would like to borrow these distribution CDs or create fresh redhat 9 CDs from their images can either contact me(sandip @ linux-delhi.org) or Yash(yash @ linux-delhi.org) offlist.
Thanks, Sandip
-- Sandip Bhattacharya sandip (at) puroga.com Puroga Technologies Pvt. Ltd. Work: http://www.puroga.com Home: http://www.sandipb.net
GPG: 51A4 6C57 4BC6 8C82 6A65 AE78 B1A1 2280 A129 0FF3
_______________________________________________ ilugd mailinglist -- [EMAIL PROTECTED] http://frodo.hserus.net/mailman/listinfo/ilugd Archives at: http://news.gmane.org/gmane.user-groups.linux.delhi http://www.mail-archive.com/[EMAIL PROTECTED]/
