Hi!

I recently downloaded iso file of CentOS 5.4. I matched the md5sum of the downloaded file with the value given at site. Both matched exactly - means there was no problem during download.

When I burnt the same iso image on DVD( checked md5sum just before burning -- it was okay )and checked md5sum again, it was found to be different. Is it normal or some problem occurred during burn process. What should I do next ?


rajesh.bahl

How did you check the md5sum of the burned dvd image? If done properly, you should get the same md5sum. However, doing so properly with optical drive media is not as simple as it may at first glance seem. See this article for a discussion of some of the issues involved:

http://www.troubleshooters.com/linux/coasterless.htm

We should get the same reply but it is not happening. I tried writing another DVD and checked the md5sum with the following:

md5sum /dev/dvd

Once again the results were different than published at site.Of course, the md5sum of both he burnt dvds were exactly same !

I just do not understand why this is happening. Also appearently, the DVDs are working okay but I have serious doubts on the integrity of packages.

Can someone help ?



rajesh.bahl

Using the rawread script given in the above link, determine the the md5 sum of the dvd by running:

$ rawread /dev/dvd | md5sum

The script uses the isoinfo command to determine the block size and block count of the cd and/or dvd. You need that info to get an accurate md5sum read.

By the way, the typical way of getting an md5sum of a dvd or cd is with this command:

$ cat /dev/dvd | md5sum

You can try that but it won't be as accurate as the rawread script method.

Finally, from personal experience, most of the problems I've run into in getting md5sums of optical media are tied to the optical drive. Some optical drives will simply not give accurate md5sums no matter what method I've used to get them. The drive works fine otherwise and the cd or dvd is good as confirmed by verified md5sums obtained on other working optical drives. I'm not sure why that is but when using the rawread script on the problem drives, you will see a lot of read error problems being experienced when doing the operation. Others I've talked to have noticed the same thing. My general theory is that when doing an md5sum from an optical drive you are accessing the drive with some pretty low level tools like cat or dd. Optical drives have built in read error correction mechanisms which I believe may not function properly when using those low level tools but the drive may work fine in normal operation.