The UNIX and Linux Forums

do a full comparison of folder contents in script

tlarkin — Fri, 20 Nov 2009 21:34:59 GMT

Hello everyone....

I have a small issue here at work and I am trying to script out a way to automate a fix for it. I have a small number of users (I work in a 1:1 with 6,000 macbooks) that aren't really managed in my deployment. They are managed with a few policies, but the policies are broken because they never turn their laptops in or get them reimaged. I have had to wipe LDAP a few times for upgrades/maintenance/fixes and their synchronized portable home directory has a different UID than the new UID since I did fresh imports of user accounts both times I destroyed and rebuilt LDAP.

So, I have a script written which seems to be going the correct way of doing this, but I am not exactly a Unix wizard so I have some basic questions hopefully someone here can help me with.

The script's function will be to copy all local user accounts in /Users to /Users/username.bu for a full back up of the account. Then, it will initiate a log out, and a log out hook that will run the second script that will verify all data is copied to the back up folder (.bu) and then delete the account from directory services and wipe the original data out of their local user account.

This way, when the user logs in next, they will have pulled down the new account to their machine that has the exact same user/password but this time it will be the right UID. Then they will be in the proper OD groups and get group policy as normal.

I hope that makes sense. Also, only accounts that have a UID higher than 1000+ are directory accounts, so I want to ignore all accounts with an UID less than 999.

So far here is my script I wrote, and like I mentioned, I am completely self taught and have no formal education in programming.

Code:

#!/bin/bash



#

# First we need to grab the synchronnized user account

# All directory accounts will have a UID greater than 1000

# All local accounts will have a UID of 500 to 999

# We only want to move accounts with UID of greater than 1000

# This will also exclude the shared folder

#



userlist=`/bin/ls /Users | /usr/bin/grep -v "^Shared"`



checkuid=`/usr/bin/dscl . read /Users/$i UniqueID | /usr/bin/awk '/ID/ { print $2 }'`



for i in $userlist ; do



if (( $checkuid > 999 )) 



    then /bin/mkdir /Users/$i.bu ;; /bin/cp -R /Users/$i /Users/$i.bu



    else /bin/echo "$i is not a directory user"



fi



done



#

# now the tricky part we need to verify what was done, is done then delete the old account

# so the user can sync the new account with the proper UID

#



homefolder=`/bin/ls -l $userlist/`



backup="/Users/$userlist.bu"



# now compare contents



if [[ $homefolder == $backup ]]



    then /usr/sbin/dscl . -delete /Users/$userlist



    else /bin/echo "Contents do not match, exiting"



    exit 1



fi



done



#

# Now remove the contents of those user accounts by comparing what has 

# .bu on the end

#



removeuser=`/bin/ls /Users | /usr/bin/grep ".bu"



/bin/rm -rf /Users/$removeuser



exit 0

Now the second script will just copy that data back after the user logs in again and resynchronizes their PHD with the proper UID. Then the last bit will apply ownership to that new UID and exit.

A few questions....

1) Should I have used an integer comparison when comparing the current UID to it being greater than 999?

2) There are probably some issues with this script as I have not tested it out fully yet, because I was wanting some feed back to see if there was a better way to go. Is there a more efficient way to go about this?

Thanks for anyone who can help, it is greatly appreciated.

Tom

Good Consultants

iBot — Fri, 20 Nov 2009 21:15:05 GMT

2009-11-20T21:58:31+01:00

Source...

help with script

dsravan — Fri, 20 Nov 2009 21:05:29 GMT

I have a file which is comma delimited. I have around 35 fields in each row. The problem is in the 21st field I have data separated by �;� I need the data in only the 21st field separated by �;� to be replaced by �|�. This is a very big file and I need something in awk that its much faster than regular commands. Can anybody tell exactly how I can only achieve this for 21st field.

How to Pull out multiple files from DB table and redirect all those files to a differetn directory?

ss3944 — Fri, 20 Nov 2009 20:54:24 GMT

Hi everyone!!

I have a database table, which has file_name as one of its fields.
Example:

File_ID File_Name Directory Size
0001 UNO_1232 /apps/opt 234
0002 UNO_1234 /apps/opt 788
0003 UNO_1235 /apps/opt 897
0004 UNO_1236 /apps/opt 568

I have to extract all the files whose size is greater than 500 and send all those files to a different directory /apps/opt/new

How do I loop through the table and send all those file_names, which match the size criteria to a different directory?

Thanks in advance..
Please do help me out, as this is very urgent for me.

finding files with unicode chars in the filename

mshallop — Fri, 20 Nov 2009 20:00:13 GMT

I'm trying to check-in a repository to svn -- but the import is failing because some files waaaay down deep in some graphics-library folder are using unicode characters in the file name - which are masked using the ls command but picked up when piping output to more:

[root@dev-www-02 emoticons]# ls -l 1914*
-rwxrwxr-x 1 apache apache 1398 Dec 9 2008 1914OdiN_Presenta_-_o_?.bmp
[root@dev-www-02 emoticons]# ls -l 1914* | more
-rwxrwxr-x 1 apache apache 1398 Dec 9 2008 1914OdiN_Presenta_-_o_�.bmp

Optimally, I'd like to be able to search and quarantine these files into a directory out of the repository tree, but I'm brickwalling :mad: trying to figure out the search string...

I've tried variants of grep '^[A-Za-z0-9]' but can't turn up the right combination.

tia...

How to read the latest file name in adirectory.

haroon_a — Fri, 20 Nov 2009 19:49:32 GMT

Hello all,

In a shell script, I need to get the name of the latest file which was created in a directory. i.e. I have the following directory.

Code:

/tmp/public/logs/

In this directory, I want to get the latest file that was created. Can you give me a hint as how can I get the latest file, based on timestamp?

Thanks in advance...

kill -3 ... where the output file goes?

haroon_a — Fri, 20 Nov 2009 18:03:53 GMT

Hi all,

I am generating the coredump of my JBoss, and by default it puts it in to a particular directory. i.e. JBOSS_HOME/. I would like this output file to be created, lets say in /tmp/dump/.

I tried the following:

Code:

kill -3 9404940>/tmp/dump/out.txt

But it created /tmp/dump/out.txt with 0 KB, and yet it created the dump file in JBOSS_HOME/ directory.

Can you please let me know how can I have kill -3 to create the core dump file in the desired location?

Thanks,

problam with diff

nirnir26 — Fri, 20 Nov 2009 18:03:14 GMT

hi, I want to compare 2 files and see if there is a difference between them
so i did:
set aa = `diff file1.txt file2.txt`
if ($aa == "") then
echo "no difference"
else
echo "different files"
endif

and i get the error msg:
if: Badly formed number.

now when im doing the ccomperation like this:
set aa
if ($aa == "") then
echo "no difference"
else
echo "different files"
endif

it works good.
why the diff doing problams??
what can i do?

Mandriva: glpi

iBot — Fri, 20 Nov 2009 18:00:05 GMT

LinuxSecurity.com: This is a maintenance and bugfix release that updates glpi to 0.72.3.

More...

Snoop perl script

Pouchie1 — Fri, 20 Nov 2009 17:02:20 GMT

Hi ,

I would like to write a perl script with the snoop command to capture packets from a specific IP address to a node (incoming packets) and packets from that node for the same session to another node and save the capture to a file. I would like my script to be able to read my IP all the time from the keyboard. I would like also to be able to determine for how how long the script should run. Let's say 60 seconds and then kill the snoop process after that period. Below is an example of the command.

snoop -o /tmp/snooper -V 128.50.1.250

let's say my snoop is called snoop1. I would like to be able to execute that way: ./snoop1 128.50.1.250 60.......where the 128.50.1.250 is the ip I want to capture the packets from and 60 the amount of time I want the script to run and then kill the snoop process.

Thanks for your help.

Technical Article: Creating Prebundled Machine Images for the Cloud

iBot — Fri, 20 Nov 2009 16:45:12 GMT

Learn how to create a custom virtual machine image prebundled with Oracle WebLogic Server 11g and Oracle Database XE, as well as the basics of creating init scripts and configuring services to run at user-defined runlevels

More...

df command always return 0%

polexx — Fri, 20 Nov 2009 16:34:30 GMT

Hi, I don't know why but about 80% of time when I execute the command
"df -v /directory" the output is always 0% used. Anyone have clue to why this is happening? Thanks.

Auditd problem

mojoman — Fri, 20 Nov 2009 15:15:55 GMT

Hi,

I have the following my logs:

Code:

Nov 20 04:02:04 mail-07 kernel: audit: audit_backlog=326 > audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: audit_lost=4272 audit_rate_limit=0 audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: backlog limit exceeded

Nov 20 04:02:04 mail-07 kernel: audit: audit_backlog=326 > audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: audit_lost=4273 audit_rate_limit=0 audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: backlog limit exceeded

Nov 20 04:02:04 mail-07 kernel: audit: audit_backlog=326 > audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: audit_lost=4274 audit_rate_limit=0 audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: backlog limit exceeded

Nov 20 04:02:04 mail-07 kernel: audit: audit_backlog=326 > audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: audit_lost=4275 audit_rate_limit=0 audit_backlog_limit=320

Nov 20 04:02:04 mail-07 kernel: audit: backlog limit exceeded

What do I do to fix this problem? Please provide a careful explanation as I am new to auditd.

Thanks

(/var/log/messages) issue

mojoman — Fri, 20 Nov 2009 14:55:17 GMT

Hi,

I found this in my logs today for vsftpd:

Code:

Nov 17 08:43:58 ftp vsftpd: Tue Nov 17 13:43:58 2009 [pid 26389] [ccm] OK LOGIN: Client "205.150.86.90"

Nov 17 08:44:49 ftp avahi-daemon[2443]: Withdrawing address record for 205.111.86.22 on eth0.

Nov 17 08:44:49 ftp avahi-daemon[2443]: Leaving mDNS multicast group on interface eth0.IPv4 with address 205.111.86.22 

Nov 17 08:44:49 ftp avahi-daemon[2443]: iface.c: interface_mdns_mcast_join() called but no local address available.

Nov 17 08:44:49 ftp avahi-daemon[2443]: Interface eth0.IPv4 no longer relevant for mDNS.

Nov 17 08:44:49 ftp avahi-daemon[2443]: Withdrawing address record for fe80::213:20ff:feaf:2fae on eth0.

Nov 17 08:44:49 ftp avahi-daemon[2443]: Leaving mDNS multicast group on interface eth0.IPv6 with address fe80::213:20ff:feaf:2fae.

Nov 17 08:44:49 ftp avahi-daemon[2443]: iface.c: interface_mdns_mcast_join() called but no local address available.

Nov 17 08:44:49 ftp avahi-daemon[2443]: Interface eth0.IPv6 no longer relevant for mDNS.

Nov 17 08:44:50 ftp kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready

Nov 17 08:44:50 ftp kernel: e100: eth0: e100_watchdog: link up, 100Mbps, full-duplex

What is going on?

Insert Text Between Html Tags Sed / Awk ?

abacus — Fri, 20 Nov 2009 14:07:19 GMT

Hello, I have searched for the title stringon this post on this site as well as on google with no luck. I find similar examples, however I am truly having a tough time resolving the issue that I am faced with.
I have multiple html files that I need to change the title on. These are simple html text files that look like this:

Input File:

Code:

oldtitle



this is a test page for this example.

Desired Output:

Code:

newtitle



this is a test page for this example.

I have tried something like this. However I think that I am over thinking this.

Code:

sed -e "s/$]*>$[^<]*$<.*$/\newtitle\2/g" file1.html > newfile.html

Please Advise,

Abacus