General :: Package Versioning And Difference Of MD5Sum On Files?
Jun 20, 2011
I noticed something a little odd I'm hoping someone can enlighten me on. I noticed in a couple of cases that a package has the proper version, but differs in two regards.
1. The package ends up with a .el4 on the end of the version for Red Hat 4.
2. The actual MD5Sum of the files the package provides differ.
An example below:
Code:
[root@RH4ES32-MCE bin]# for i in `rpm -ql GConf2`;do md5sum $i;done;
md5sum: /etc/gconf/2: Is a directory
9f90335546f7c57ae6fb552cc2b919c5 /etc/gconf/2/path
md5sum: /etc/gconf/gconf.xml.defaults: Is a directory
[code].....
So my package changed slightly to now show .el4 versus just 2-2.8.1-1 I've indicated in the first output above that the first couple of lines differ. I stopped my comparison at that point as they truly are different.
What i am trying is to check the file duplication in a folder and remove a file if it is a duplicate of another file ie the contents are duplicate; but names may be same.
Basically i am using md5sum to calculate the md5sum values of each file and redirecting to a file. And i am thinking of comparing the md5sum values.But i am finding it hard to decide how to complete the code after redirecting the output of calculation of md5sum to a file.
I want to move all files and directories that are 1 month old out to back up into a separate folder. There will be a lot of files and I want to make sure it copies properly. The problem I'm having is integrating a MD5SUM into it to check integrity. MD5SUM is not recursive, so I figured it would work in a loop when it copies each individual file, I'll do a md5sum on each file and delete that md5 once its verified it copied ok.
[Code]...
I also need some sort of error handling to output all md5's that didnt pass the hash check.
I'm looking for a fast way to verify a copy of a folder with 150Gigs of data, in 33 files. Some of the files are a few kb, while a few are 20-30Gigs. I've done a file count, which is quick, but doesn't verify that all the files are intact. I tried running md5sum on them, which works, but will probably take as long as copying the files in the first place. Diff works too, but is slow too.
In order to upgrade a machine that can not successfully upgrade to 10.4 I downloaded and burned the 10.04.1 iso image off the ubuntu alternate download site. In my first attempt I unsuccessfully burned the image with it failing at the very end. I did perform an md5sum on it and received the precise output I got from my second burn attenpt which DID complete successfully. Here is the output:
[code]...
I did research this last night and it seems the common wisdom was to reburn the iso (which I did twice) or copy down the iso again. This I also did and it came down precisely, bit for bit, the same as the first one. Here are the two cksums
[code]...
Is there something wrong with this image on the website or is the error about 1 file being unreadable (could that also mean missing?) be erroneous?
What is the difference between "bin" and "netinstall" ISO files for Linux?I need a full OS, not just a Live CD. On the mirrors there are two kind of files.Which one is the full OS?
I have a Centos 5.5 server and I had a problem with its time because it was 1 one hour ahead of my local time. I installed and activated NTP and I created a link for /etc/localtime:
I reboot the server and waited for 1 hour but the time wasnt correct. The server's BIOS clock has UTC time so I edited /etc/sysconfig/clock file replacing:
Code:
UTC=false for UTC=true
I reboot the server and waited for 1 hour but the time wasnt correct. After breaking my head for about 1 hour, I realized that there is a directory /usr/share/zoneinfo/Mexico and changed my link for /etc/localtime
And It worked. What is the difference between America/Mexico_City and Mexico/General files in /usr/share/timezones, they should be identical but of course they are not?
Am just wondering what the notion is behind having a package without all needed files, and having a -devel package with the rest. Such that whenever am installing a package fron tarball, and it says it cannot get a package, I always install the -devel equivalent and all gets to be OK. Why dont they just have everything in the original package?
I have some checksums.md5 verification files from an ntfs external drive, but using windows notation: instead of /, spaces between file names (not escaped), reserved shell characters (like (, &, ', to name a few). The checksums.md5 has a bunch of checksums and filenames:
[code]...
I want to use this checksums.md5 to verify the files that I've copied to my machine: but I'm on a Linux, so I need to convert the names inside checksums.md5 from Windows to Linux to use the md5sum utility from the shell. The first line in my example would become: f12f75c1f2d1a658dc32ca6ef9ef3ffc My Windows & Files (2010)/[bak]/testing.wmv Is there some application for this (converting a file listing, from windows cmd notation, to linux shell notation) or will I need to create a bash script using sed that just "replaces" what is "wrong" with the filenames
OS: Debian testing Is it possible to instruct apt-get to install only some of the files in a package?.Here is the specific problem I'm facing: I have installed latex using the latex installer from CTAN. Now I'd like to install lyx but when apt-get install lyx reports it will install all the deb latex packages (it seems to be unaware latex is already installed). can I instruct apt-get to install only the lyx files with the extra 300MB or so of latex?.
Where is the best path to install all files of package when compiling?
For example - I want to install ProFTPd, so there's an option
--prefix=/usr/local/proftpd
, which means, that all files after compilation (including binary and configuration files) will be stored here. As you know, all packages, which are installed via package system (like zypper on SuSE or apt on Ubuntu) usually stores its configuration files in /etc/ and binary files in /sbin/ and also stores a link in my $PATH, so i can run in just by typing proftpd (without /sbin).
The huge advantage of packages installed via package system, i think, is easy uninstall process. I'd like to be able to uninstall compiled packages also easy.
I think i could be able to create some batch file, which I'll be able to use like this:
uninstall --package=proftpd
And my script will find all proftpd files in usuall paths (/etc, /sbin) and remove it using rm.
Is there any best practices, where to store all these files, or are there any (dis)advantages of my first example (--prefix=/usr/local/proftpd)?
I really don't think, that it's nice to have 2 paths with configuration files and binary files, but maybe I'm misunderstanding the basic principals of Linux... :-)
i am using diff command... to get difference between two file. but the thing is its giving both file difference and i need only the difference of 1st file comparing to 2nd file for that it shoud not show anything abt 2nd file.
1)Say an application is written and is having source files. 2)A file called source rpm package 3)A rpm package
In case of (1) we need to compile using either the make or if we know using appropriate compiler with proper libs included. Can I know what's the difference between (2) and (3)?
There is .profile under my directory and I found another one under /etc/ . I know the first one is the one we edit for path and alias and stuff. but how does it differ from the one in /etc?
I just upgraded my kernel to version 2.6.38 & I seem to have missed a few modules. I tried using diff to get a list of the differences between the files lsmod-2.6.35 & lsmod-2.5.38, which I created for this. Unfortunatly I can't seem to figure out the right syntax to do what I want which is to just show me whats missing. For example, I want to subtract the lines from file1 & file2 which are the same, leaving me with the difference in file3. i.e. file3 = file1 - file2 Is this possible or am I just using the wrong program.
I've just installed ubuntu 10.04 and see that there are a two type installation software? why there are two things? What are the benefits of these two? What are the Difference Between Synaptic Package Manager and Update Manager?
However, I noticed that there are some difference in the headers files that get generated with the above command as compared to the header files that are available from Debian repo. For example, the header files for 2.6.24 kernel have files like:
Code:
debian:/usr/src/linux-headers-2.6.24-etchnhalf.1-686/arch/x86# ls Kconfig Kconfig.cpu Kconfig.debug kernel Makefile Makefile_32 Makefile_32.cpu Makefile_64 However, if I custom compile the above kernel from Debian sources (2.6.24), the headers files does not have the above files:
Code:
debain:/usr/src/linux-headers-2.6.24-generic-ide/arch/x86# ls boot ia32 Kconfig.cpu kernel lib mach-es7000 mach-visws Makefile mm pci vdso xen crypto Kconfig Kconfig.debug lguest mach-default mach-generic mach-voyager math-emu oprofile power video
As you can see from above output, files like Makefile_32, Makefile_32.cpu are not present, if I generate kernel headers files using the make-kpkg command as mentioned in the beginning of the post. I happened to notice the above issue, while I was trying to compile a out-of-tree kernel module and the "make" command for those sources (some graphics card drm module) worked with the default header files (linux-headers-2.6.24-etchnhalf.1-686) but did not work with (linux-headers-2.6.24-generic-ide) because it did not find the Makefile_32. Although I was able to fix the problem by copying the Makefile_32 from linux-headers-2.6.24-etchnhalf.1-686 but I would like to know why there is a difference. This is bit of a concern because it unnecessarily breaks the out-of-tree module compilation process because of trivial reasons.
How can I recover My deleted files in ubuntu? What's the difference between "foremost" and "scalpel"? And is there any other program(or package?) For this purpose in ubuntu? I am running ubuntu 9.10
