david.rabine

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Backups ======

===== Account Migration =====

==== Find User Files ====

Find all the files that a specific user owns. (and ignoring remote mounts)  Need to sudo or be root to do this.  **-xdev** will keep the find to the local hard drive so it doesn't search remotely mounted drives.

<code>
find / -type f -user david -xdev > ~/david_files.txt
</code>

And, if you would like to see what files in this list are not in your home directory, try this:

<code>
cat ~/david_files.txt | grep -v '/home/david'  # or you could use $HOME if you like in place of '/home/david' 8-)
</code>

===== Checksum Directory Contents =====

==== Compare Code Folders ====

I am comparing a local drive versus a checked out repository and these steps can validate they match.  Using the find command will recursively dig down to find all the relevant files.

  - Create a listing of the files and md5sum of one of the directories.  Here I am putting together all the .cpp and all the .h files into a single listing <code>
# create md5deep of current repository code

rm ~/gui_src_md5deep.txt
find lvisf-gui/ -iname '*.cpp' -exec md5deep -l {} \; >> ~/gui_src_md5deep.txt
find lvisf-gui/ -iname '*.h' -exec md5deep -l {} \; >> ~/gui_src_md5deep.txt
</code>
  - Compare all the .cpp and .h files in the other directory against this list.  No return should happen with either of the commands.  If you get anything back, it doesn't match.  This is a little off putting to get no feedback, so if you like... modify one of the md5sums and see it fail and show you it doesn't work. <code>
# compare to what is on the system

cd /home/lvis/work/lvisf-gui
ln -s trunk lvisf-gui
find lvisf-gui/ -iname '*.cpp' -exec md5deep -lX ~/gui_src_md5deep.txt {} \;
find lvisf-gui/ -iname '*.h' -exec md5deep -lX ~/gui_src_md5deep.txt {} \;
</code>

==== Copy Large Disk ====

If you are in need of moving a large amount of files & links from one drive (raid probably) to another.  This would be one way to go about that.

  - Make note of total size, file count, directory count and symbolic link count of the original data <code>
# define our data directory
export MYDATA='/space/data/'

# total size of partition
du -s $MYDATA
4823313734      data

# directory count
find $MYDATA -type f -iname '*' -print | wc -l
3632

# file count
find $MYDATA -type d -iname '*' -print | wc -l
305

# symbolic link count
find $MYDATA -type l -iname '*' -print | wc -l
3

# make a list of symbolic links you can diff later on (column 9 might be wrong for your OS, check)
find $MYDATA -type l -iname '*' -print | xargs ls -l | sort -k 9 > ~/mybackup_symbolic_link_details.txt

</code>
  - Generate a checksum manifest for the data (just files, not symbolic links) <code>
time md5deep -o f -rlz $MYDATA | sort -k 3 > ~/mybackup_md5deep_all_files.txt
</code>
  - Copy the data (I still like rsync) <code>
time rsync -vaP $MYDATA /to/your/destination
</code>
  - Wait a long time
  - Check that the values match <code>
export NEWCOPY='/mnt/tmp/newraid/data'
du -s $NEWCOPY
find $NEWCOPY -type f -iname '*' -print | wc -l
find $NEWCOPY -type d -iname '*' -print | wc -l
find $NEWCOPY -type l -iname '*' -print | wc -l
</code>
  - Create a symbolic link inventory and diff the results <code>
find $NEWCOPY -type l -iname '*' -print | xargs ls -l | sort -k 9 > ~/mynewdrive_symbolic_link_details.txt
# if this does not work... uh oh.
diff ~/mybackup_symbolic_link_details.txt ~/mynewdrive_symbolic_link_details.txt
</code>
  - Verify md5deep of files (and you should get NOTHING back) (could pipe this to a file) <code>
time find $NEWCOPY -type f | xargs md5deep -X ~/mybackup_md5deep_all_files.txt

# piping to a file
time find $NEWCOPY -type f | xargs md5deep -X ~/mybackup_md5deep_all_files.txt > ~/mynewdrive_md5deep_mismatches.txt

# and the mismatch file size should be zero (-8
ls -l ~/mynewdrive_md5deep_mismatches.txt
</code>
==== Create md5sum List ====

Using **md5sum** or **md5deep** to create md5sum lists for each directory (as borrowed from here: [[http://www.linuxquestions.org/questions/linux-software-2/how-to-create-md5sum-for-a-directory-689242/]])

  - use find in conjunction with md5sum:<code>
find directory -type f -print0 | xargs -0 md5sum >> checksums.md5
</code>
  - go to each directory, and use md5sum to check the files:<code>
md5sum -c /path/to/file.md5
</code>

**OR**

  - Use recursive md5deep:
<code>
md5deep -rl directory > checksums.md5
</code>

==== Verify Files ====

Once you have a md5sum list, you can verify the files using **md5deep** (and this input list doesn't have to be in any order, just verifies the md5sum and file name):

<code>
# the -X shows you only files that do not pass, try -M if you want to see all the successes
md5deep -X /PATH/TO/CHECKSUMS.md5 /DIR/TO/CHECK/*
</code>

===== Mirror a Hard Drive =====

**dd** is your best friend here.  This page explains most anything you would want to do:  

  * [[http://www.linuxquestions.org/linux/answers/Applications_GUI_Multimedia/How_To_Do_Eveything_With_DD]]
  * [[http://sleepyhead.de/howto/?href=other]] <- Good linux notes / tips

==== Mirror a Drive Partition to a File ====

  - Connect the drive to your system, but it **MUST BE UNMOUNTED** when you do this.  An open / active partition when mirrored can spawn an unstable replication of itself.
  - Use **dd** to pipe the partition to your file:  <code>
dd if=/dev/sdX1 of=/some/path/to/output/to.img bs=4096 conv=notrunc,noerror
</code>

==== Mirror the Entire Drive ====

NOTE:  Be VERY CAREFUL that you're writing to the drive you want to write to.   A simple mistake will wipe the entire wrong drive and is SO easy to do.  **if** is IN FILE and **of** is OUT FILE

I added the **time** command to see how long it takes to do this drive mirror.

<code>
time dd if=/dev/sdX of=/dev/sdY bs=4096 conv=notrunc,noerror
</code>

If you have bad sectors and want to "pad" and bad data with zeros (where it cannot read data from the source disk), use the **sync** option:  (block size ok?  old disks were 512, I believe 4096 is correct currently (2012))

<code>
time dd if=/dev/sdX of=/dev/sdY bs=4096 conv=notrunc,noerror,sync
</code>

Updated today using lvis classic OS drive as an example  --- //[[david@lattice.net|David Lloyd Rabine]] 2019/02/08 14:39//
<code>
# compress the entire disk image into a compressed file:
time sudo bash -c "dd if=/dev/sdc bs=512 conv=notrunc,noerror,sync | gzip --best > /tmp/lvisclassic_os_dd_bs512.img.gz"

time sudo bash -c "dd if=/dev/sdd bs=4096 conv=notrunc,noerror,sync | gzip > /media/ubuntu/LVIS-EXT3-2/lvisclassic_os_dd_bs4096.img.gz"

#32 GB 2021.07.13
time dd if=/dev/sdd conv=notrunc,noerror | pv -s 32017047552 | pigz --fast > /media/ubuntu/LVIS-EXT3-2/lvisclassic_os_dd_pigz_nobs.img.gz

# md5deep check the value of the raw data
time gunzip lvisclassic_os_dd_pigz_nobs.img.gz -c | pv -s 32017047552 | md5deep -z
gunzip lvisclassic_os_dd_bs512.img.gz -c | md5deep -z

mv gunzip lvisclassic_os_dd_bs512.img.gz lvisclassic_os_dd_bs512.img.20190208.a7c58c89e09befd7b6b871606cfb4821.gz

32017047552  a7c58c89e09befd7b6b871606cfb4821

# output back to a device

gunzip lvisclassic_os_dd_bs512.img.20190208.a7c58c89e09befd7b6b871606cfb4821.gz -c | dd of=/tmp/diskimage.img bs=512 conv=notrunc,noerror,sync status=progress
</code>

Add a progress bar and use **sudo**
<code>
time sudo dd if=/dev/sdX of=/dev/sdY bs=4096 conv=notrunc,noerror status=progress
</code>

Latest attempt at mirroring Sylvia's laptop (2019.12.27) she got a new 1TB SSD

<code>
time sudo dd if=/dev/sdX of=/dev/sdY bs=64k conv=notrunc,noerror status=progress

kubuntu@kubuntu:~$ time sudo dd if=/dev/sda of=/dev/sdc bs=64K conv=notrunc,noerror status=progress
1000203550720 bytes (1.0 TB, 932 GiB) copied, 11060 s, 90.4 MB/s 
15261915+1 records in
15261915+1 records out
1000204886016 bytes (1.0 TB, 932 GiB) copied, 11062.5 s, 90.4 MB/s

real    184m22.400s
user    0m29.380s
sys     33m49.240s
</code>

==== Using tar Compress ====

=== Specific File Types ===

<code>
find /path/to/directory -type f -name "*.extension" -print0 | tar -cvf archive_name.tar --null -T -
</code>


Add compression

<code>
find /path/to/directory -type f -name "*.txt" -print0 | tar -cv --null -T - | pigz --best > archive_name.tar.gz
</code>
====== BIOS Upgrade ======

===== USB Boot =====

A machine at work had no CD and no floppy, but I needed a BIOS update.  I used a linux machine to write a free DOS image to my 64MB USB stick and that allowed the machine to boot and left enough room for the flash image and updating software.  Worked great:  

  - I downloaded the FREEDOS image from this page:  [[http://www.fdos.org/bootdisks/|FREEDOS Boot Disks]]
    * Specific image used:  [[http://www.fdos.org/bootdisks/autogen/FDSTD.288.gz]]
  - Insert the USB stick (if it comes up and gets mounted, you'll want to umount it:  just right click on it and unmount)
  - Write the image to the USB stick (this was on an Ubuntu machine, so use **sudo**):  <code>
##### WARNING:  this formats the drive completely, everything on that drive WILL BE LOST #####
sudo dd if=FDSTD.288 of=/dev/sdxxx bs=512  # replace sdxxx with your machine's USB
##### (BE VERY CAREFUL which disk you are wiping out) #####
sync # doubt you need to do this, but why not...
</code>
  - Re-insert the USB, and copy the files you need for the BIOS update to the USB stick
  - Unmount the USB stick and insert into your machine for booting / BIOS update
  - For the Tyan board, I had to use the following procedure:
    - Insert the USB stick while the system is powered off
    - Power up the system
    - Enter the BIOS (F2), head to the boot section and make the USB stick the first booting device
    - Save changes and reboot

====== CD ======

===== Burn A CD =====

If you have an ISO image to burn, this should work (provided your drive is a write-able one)

<code bash>
cdrecord dev=/dev/hdXX slackware-11.0-install-d1.iso
</code>

or something like this if you want to control the speed and see verbose messages

   cdrecord -v speed=4 dev=/dev/hdXX slackware-13.0-install-d1.iso

====== DVD ======





===== Create From AVI =====

  - [[http://gentoo-wiki.com/HOWTO_Create_a_DVD:Encode|Encode]] your video.
  - [[http://gentoo-wiki.com/HOWTO_Create_a_DVD:Filesystem|Generate]] the VOB files.
  - Burn it to a DVD with Nero or something.

===== Create From MythTV Stream =====


  - This [[http://tsformer.berlios.de/|Transport stream transformer]] program should be able to take an over the air stream and allow you to record it to a DVD.

===== Verify Image =====

Here is a good link for [[http://www.g-loaded.eu/2006/10/07/verify-a-burned-cddvd-image-on-linux/|md5sum checking an ISO]].

Trying to burn all these MST3k DAP-DVDs, I wanted to ensure that the burned disc was exactly the file.  So, this little script should do it.
(OK, it ended up being not so simple/little, but I think this is a solid solution.  My only worry is if the cat wildcard, the files don't sort properly?)

**verify_mst3k_dap_dvd.sh**
<code bash>
#!/bin/sh
#
# verify_image:  Input an mds file, and verify the image with the drive
if [ -z "$1" ]; then
  echo
  echo "usage: $0 <FILE.mds> [dvd drive]"
  echo "dvd drive example:    /dev/hdb"
  echo
  exit
fi
SRCMDS=$1
DVD=$2
if [ -z "$2" ]; then
  DVD=/dev/hdc
fi

echo "Verifying $SRCMDS matches with DVD in $DVD"
FILEROOT=${SRCMDS%.mds}
echo "File root = $FILEROOT"
NOOFFILE=`ls "$FILEROOT.i"* | wc -l`
FULLFILES=`expr $NOOFFILE - 1`
# get the size of the files, but ignore any that are FULL (this could break if a disc was completely full?)
SMALLBYTES=`du --bytes "$FILEROOT".i* | grep -v 734003200 | awk '{ print $1}'`
if [ -z "$SMALLBYTES" ]; then
  # init the value with the default if we got nothing from the inverse grep
  SMALLBYTES=734003200
fi
FULLBLOCKS=`expr $FULLFILES \* 358400 + $SMALLBYTES \/ 2048`
echo "Running MD5sum of disk (block size=2048 count=$FULLBLOCKS)..."
MD5DVD=`dd if=$DVD bs=2048 count=$FULLBLOCKS | nice -n 19 md5sum | awk '{ print substr($1,0,32) }'`
echo "$MD5DVD is md5sum of $DVD"
echo "Running MD5sum of files...."
MD5FILES=`cat "$FILEROOT".i* | nice -n 19 md5sum | awk '{ print substr($1,0,32) }'`
echo "$MD5FILES is md5sum of files"
if [ "$MD5DVD" != "$MD5FILES" ]; then
  echo "ERROR:  MD5sums DO NOT MATCH!?!"
  exit 1
else
  echo "SUCCESS:  MD5sum of DVD matches files."
  exit 0
fi
</code>

====== Hard Drive ======

===== Test Disk Write Speed =====

I found this [[http://it.toolbox.com/blogs/database-soup/testing-disk-speed-the-dd-test-31069|here]], and will just replicate it for consolidation.

<code bash>
time sh -c "dd if=/dev/zero of=ddfile bs=8k count=2000000 && sync"
</code>


====== MYSQL ======

===== Direct Commands =====

==== Replace ====

Here's a great trick if you want to replace all occurances in a table of something with something else.  This is an example I did at work to update the directories in our Joomla Remository:

<code>
UPDATE mos_downloads_files SET filepath = 
replace(filepath,"/web/lvis/downloads/data_sets/LDS_1.02/","/data/www/lvis/downloads/data_sets/LDS_1.02/");
</code>

====== Networking ======

===== rsync limiting bandwidth =====

If you don't want to be **that guy** and eat up your entire network's bandwidth.  You can limit the speed of a transfer say with **rsync** by including a bandwidth limit option.

Here is an example keeping the speed to 75 mbps (half my current speed) (examples are from [[https://www.cyberciti.biz/faq/how-to-set-keep-rsync-from-using-all-your-bandwidth-on-linux-unix/]])

<code>
rsync --bwlimit=KBPS src dst
rsync --bwlimit=KBPS [options] src dst
rsync --bwlimit=KBPS [options] src user@dst
rsync --bwlimit=KBPS [options] user@src /path/to/dir
rsync --bwlimit=KBPS -a -P /path/to/dir/ user@server1.cyberciti.biz
rsync --bwlimit=7500 -a -P /path/to/dir/ user@server1.cyberciti.biz # 7.5 MBytes/sec
</code>

===== rsync through tar =====

This works on the mac much better than pure **rsync** (found here:  [[https://unix.stackexchange.com/questions/30953/tar-rsync-untar-any-speed-benefit-over-just-rsync]])

<code>
tar -C /src/dir -jcf - ./ | ssh user@server 'tar -C /dest/dir -jxf -'
#
# example (doesn't make the directory like rsync does)
#
time tar -C /local/stuff -jcf - ./ | ssh username@machinename 'tar -C /Volumes/blah/incoming/remote/stuff -jxf -'
#
# do not compress
#
time tar -C /local/stuff -cf - ./ | ssh username@machinename 'tar -C /Volumes/blah/incoming/remote/stuff -xf -'

</code>

Here is how you can pipe through **tar** through **pigz** (on this machine, I had to give the full path to pigz, just FYI)

<code>
# can also add --exclude="" on the first tar line to exclude a subdir
cd /mnt/external/some/file/directory/
time tar -cf - filename.bin | pigz | ssh user@server "/usr/local/bin/pigz -d | tar xf - -C /destination/path/"
</code>

copy FROM a server back to yourself 
<code>
bash-3.2$ ssh username@server "tar czf - --strip-components=PATHDEPTH /server/source/directory" | tar xzf - -C /your/local/dest/
</code>

===== rsync Windoze Files =====

You need to expand the time stamp difference for Windoze to not re-copy everything;

<code>
# always have vP for verbose and progress
rsync --modify-window=5 -vPrltD NORMAL THING
# NO LUCK for me like this, but adding this does only check file size

rsync --size-only -vaP FROM/ TO/

# try this
rsync --size-only -vPrltD # OR remove -t

# doing this on windows backup drive:
rsync --size-only -vPrlD 
</code>
===== sshd Piping =====

In order to access something through a machine, you could pipe http (or anything) over **sshd** with something like this:

  - Open a ssh connection with the -L switch, redirect incoming **ssh** pipes to the port / machine you want: (and -C for compression)<code>
ssh -C -L 8000:localhost:80 david@avalon.gsfc.nasa.gov
</code>
  - Then go to your local host with your browser and you should be piping over to port 8000 on avalon (or click [[http://localhost:8000|here!]]): <code>
URL:   http://localhost:8000
</code>

===== sshd Tunnel =====

==== Working ====

NOTE: Verified today this works from South Africa between my servers.    --- //[[david@lattice.net|David Lloyd Rabine]] 2023/10/19 10:07//

On the machine you want to get connected to: (-N to not get a prompt, just create the tunnel)

  ssh -R 6333:localhost:22 username@youroutsideserver.com
  # or
  ssh -N -R 6333:localhost:22 username@youroutsideserver.com
  

log into that machine and then connect to your box with:

  ssh -p 6333 localhost
  # or
  ssh -p 6333 userOnAboveMachine@localhost

==== Richards Solution ====


Keeping a sshd tunnel open from a hotel connection to a trusted system.  Richard suggested doing something like this:

Here is an example:  **lifeline.sh** (run this on the system from the hotel to open up a tunnel to your trusted computer)
<code>
#!/bin/sh
# set and keep up a tunnel to our trusted system
while ( /bin/true ); do
  ssh -CX -R 5961:localhost:5901 -R 2232:localhost:22 trusted.net
done
</code>

This setup then allows remote workers into our systems at the hotel to process data.  Adjust the port numbers as necessary and run the code on any system someone temporarily gives you access

====== Photos ======

===== Resize =====

If you want to resize a bunch of files with linux, try this following command line I found on [[http://www.perturb.org/display/ImageMagick_resize_images.html|this webpage]] (uses the ImageMagick suite)

<code bash>
for i in *.jpg; do convert -resize 50% -quality 80 $i conv_$i ; done
</code>

====== Raid ======

===== btrfs =====

This looks like a VERY flexible file system where you can add devices after the FS is created, so expansion is infinite without having to reformat, etc...  Pretty neat.  What happens if you remove a device and there isn't enough space left to hold what is on there!?  I am pleasantly confused!  8-)

  * Munchy sent me this link, proclaiming the death of raid!  I think it is raid, just completely open / wild:  [[https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices]]

===== Linux Software Raid =====

I'm setting up my little portable raid-0 drive as a linux software raid 1 so if a disk fails, it will just be rebuilt under linux.

A lot of this is coming directly from [[http://ubuntuforums.org/showthread.php?t=408461|here]] and the man page for **mdadm**.

  - Attach the external drive via firewire, as for whatever reason in USB it only sees one drive.
  - Create a partition on each drive (just take up the entire disk) with **fdisk** on each.
  - Use **mdadm** to create the array <code>
mdadm --create /dev/md0 --chunk=64 --level=1  --name=drabineraid --raid-devices=2 /dev/sda1 /dev/sdb1
# on chewy, i did this (since it already had a couple SCSI drives)
mdadm --create /dev/md0 --chunk=64 --level=1  --name=drabineraid --raid-devices=2 /dev/sdd1 /dev/sde1
</code>
  - <del>Autobuild the array using **mdadm** </del> (Hmmm, said already active (I guess the build activated it) )<code>
mdadm --assemble /dev/md0 /dev/sdd1 /dev/sde1
# or after you have a config file
mdadm --assemble --scan
</code>
  - Format the raid partition as ext3 <code>
mkfs.ext3 -m 1 /dev/md0 -L rabineraid # only do a 1 percent reserve for the super user (this is my user space) and label as 'rabineraid'
</code>


To stop the software raid: <code>
mdadm --stop /dev/md0
</code>

==== Raid Status ====

=== Automated Monitoring ===

Put this in your **/etc/rc.d/rc.local** to automatically monitor your arrays and email yourself the results:
<code>
mdadm --daemonise --monitor --mail YOUR.EMAIL@YOUR.SERVER.COM --test /dev/md0
</code>


=== Command Line Details ===

Here is an example of querying the status of the raid while it is building up:

<code>
[root@gauss ~]#  mdadm --detail /dev/md0
/dev/md0:
        Version : 00.90.03
  Creation Time : Mon Jan 14 13:50:16 2008
     Raid Level : raid1
     Array Size : 488383936 (465.76 GiB 500.11 GB)
    Device Size : 488383936 (465.76 GiB 500.11 GB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Mon Jan 14 13:50:16 2008
          State : clean, resyncing
 Active Devices : 2
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 0

 Rebuild Status : 0% complete

           UUID : 74b7c532:2225ed7e:4fa0d22e:5291ec50
         Events : 0.1

    Number   Major   Minor   RaidDevice State
       0       8       49        0      active sync   /dev/sdd1
       1       8       65        1      active sync   /dev/sde1
</code>

====== Repositories ======

===== git =====

Finally moving into the 21st century and using **git** as our repository for a lidar project at work.

  * Be sure and set your name and email variables before doing this <code>
git config --global user.name "YOUR NAME HERE"
git config --global user.email "YOUR.EMAIL@YOUR.EMAIL.HOST"
</code>

Not sure what this does, but Joey says "after/before you switch to the other branch"  <code>git fetch</code>

  * Cloning to your local folder <code>
git clone git@aetd-git.gsfc.nasa.gov:joseph.e.gurganus/splice-fsw.git 
</code>
  * Adding items to the repository
    * Tag new items as things to add <code>
git add .
</code>
    * Commit the stuff <code>
git commit -m "Some comment probably a good idea"
</code>
    * Push the update to the repository <code>
git push
</code>
  * Updating files
    * Make your changes
    * Make sure you're in the correct branch 8-)
    * Add the files you want to update "when you change a file which is already in the repository, you have to **git add** it again if you want it to be staged." [[https://stackoverflow.com/questions/21134960/what-does-changes-not-staged-for-commit-mean]]<code>
git add a
git commit a -m "bugfix, in a"
git add b
git commit b -m "new feature, in b"
</code>
  * Updating the master branch from the repository <code>
# to update the dev branch
git pull origin dev

# to update the master
git pull origin master
</code>
  * Creating a branch
    * Check existing branch information <code>
 git branch -ra
</code>
    * Switch to **dev** branch for example <code>
git checkout -b dev
</code>
    * What branch am I in? <code>
git branch
</code>
    * Push the code <code>
git push # probably error
</code>
    * Push this way to have it track properly (am I saying this right?) <code>
git push --set-upstream origin dev_dlr
</code>

  * Using Git to checkout a branch on the command line [[https://confluence.atlassian.com/bitbucket/checkout-a-branch-into-a-local-repository-313466957.html]]
    * Change to the root of the local repository. <code>
cd <repo_name>
</code>
    * List all your branches: <code>
$ git branch -a
You should see something similar to the following:

* master   <feature_branch>
  remotes/origin/<feature_branch>
  remotes/origin/master
</code>
    * Checkout the branch you want to use and pull it down. <code>
$ git checkout <feature_branch>
$ git pull
</code>
    * Confirm you are now working on that branch: <code>
$ git branch
</code>
      * You should see something similar to the following: <code>
$ git branch
* <feature_branch>
  master
</code>
===== SVN =====

==== Revision Differences ====

If you want a list of which files were changed during a check-in (for instance... if that broke everything and you just want to go back to the revision before it).  Use this command:

  svn diff -r 167:168 --summarize


====== Shell Commands ======

===== Compression =====

==== Multi-threaded Tar ====

[[https://www.cnx-software.com/2016/12/16/compress-decompress-files-faster-with-lbzip2-multi-threaded-version-of-bzip2/]]

An example of using **lbzip2** or **pigz** to multi-threaded decompress the tar file. 

<code>
time tar tvf backup_work.tar.bz2 --use-compress-program=lbzip2

time tar tvf backup_file.tar.gz --use-compress-program="pigz --processes 8" > backup_file.tar.gz.ls.txt

time tar cf - paths-to-archive | pigz > archive.tar.gz

tar cf - paths-to-archive | pigz -9 -p 32 > archive.tar.gz

</code>

==== Un-Tar a File Removing Path ====

This is nice if you are just wanting to grab a file and not leave some directories laying around that you'll need to clean up after getting that one file you really want.

<code>
time tar --strip-components 3 -zxvf SOMEARCHIVE.tar.gz path1/path2/path3/SOMEFILE
</code>
===== bash =====

==== Combine PDF Files ====

How do you combine PDF files into one big file?  I'm glad you asked!  I found this very useful for pulling my FAA documentation together from over 200 separate PDF prints from my CAD program.  I found the answer [[http://www.brighthub.com/computing/linux/articles/14795.aspx|here]]

<code>
alias pdfmerge='gs -dBATCH -dNOPAUSE -q -sDEVICE=pdfwrite -sOutputFile=./output.pdf'
</code>

Then to combine the files, just run: (and you'll end up with an output.pdf where you are)
<code>
pdfmerge FILE1 FILE2 ... etc...
</code>

==== Remove Directory Or Path ====

Found this tip [[http://stackoverflow.com/questions/125281/how-do-i-remove-the-file-suffix-and-path-portion-from-a-path-string-in-bash|here]]

  mystring=/foo/fizzbuzz.bar
  echo basename: $(basename $mystring)
  echo basename + remove .bar: $(basename $mystring .bar)
  echo dirname: $(dirname $mystring)

==== Remove Control-M Characters ====

If you have pesky control-M characters in your file, edit with **vi** and use:

  :%s/^M//g

==== Search for a File ====

If you know a part of the name or a couple parts of the file, you can **find** it with the following examples.

  * Note **iname** makes the name search not caps sensitive
  * Note **grep** can just keep on stacking if you have more and more key words to search for
  * Note **grep** is also case insensitive if you add the **-i** or **--ignore-case** option

Just copy and paste any of these commands, these all work 8-)
<code>
# say you are looking for word world (an old cartoon I captured over the air with the mythtv for the kids)

# search for world
find /pool/media/video -iname '*world*'

# but you might get a lot of hits for world that is a popular name, so use grep to add word
find /pool/media/video -iname '*world*' | grep -i 'word'

# and just to add another key word as an example let us only show mpg files
find /pool/media/video -iname '*world*' | grep -i 'word' | grep -i 'mpg'

# and if you want to clear the screen first for easy reading, stack that command
clear ; find /pool/media/video -iname '*world*' | grep -i 'word' | grep -i 'mpg'

</code>

==== Show Files by Access Time ====

Here are a few command line examples to show files by access time (always wanted a good way to do this) so you can see what the latest files added to a directory or drive are:

<code>
find /pool/media/video/hd -xdev -type f -exec stat --format=%Y%n {} \; | sort

# make it look nicer (the time is "seconds since Jan 1 1970") (n = numberic sort)
find /pool/media/video/hd -xdev -type f -exec stat --format="%Y %n" {} \; | sort -k 1,1n

# pipe to less if you want to page up / page down have a look (and reverse to so newest at top) q to exit less
find /pool/media/video/hd -xdev -type f -exec stat --format="%Y %n" {} \; | sort -k 1,1nr | less

# AI suggested this for nice file formatting 2025
# Alternative using printf for better formatting
find /path/to/search -type f -printf '%T+ %p\n' | sort -r | head -10

</code>
==== Find and Copy Files Between Two Dates ====

Use **newermt** switch to **find** and copy files that are human readable times / dates from one place to another.

<code>
find /Volumes/6608 -newermt "2024-11-01 00:00:00" ! -newermt "2024-12-01 00:00:00" -exec cp -a "{}" .  \;
</code>



==== Sort Text File By Second Column ====

If you want to sort (say a md5deep list) of files by their path instead of the md5sum itself, try this:

  sort +1 -2 md5deeps.txt

==== Where Has All My Disk Space Gone? ====

If you want to show your directories by size (to help you figure out why you have very little hard drive space left), I like to do a limited find and then use **du** to see who the offending directory is.

Here I am trying to figure out why my windows partition is full:

  sudo find /windows/ -maxdepth 1 -type d -exec du -hs {} \;

Or if you would like to sort by numeric size:

  sudo find /windows/Users/david -maxdepth 1 -type d -exec du -s {} \; | sort -n -k 1

===== copy =====

  * From:  [[https://unix.stackexchange.com/questions/30953/tar-rsync-untar-any-speed-benefit-over-just-rsync]]

Sometimes this works faster than straight up **rsync** when you pipe the ssh copy with tar (Mac OS X back in the day)

<code>
tar -C /src/dir -jcf - ./ | ssh user@server 'tar -C /dest/dir -jxf -'
</code>

===== daemons =====


==== teamspeak server ====

Here is how to start the team speak server on archon.

  - Log into archon as quake user.
  - Run the following:  <code>
cd ~/teamspeak/tss2_rc2/
./server_linux start
</code>

**NOTE:** you can replace start with either:   **start**, **stop** or **restart** (all three should be valid options)

====== Shell Scripts ======

===== bash =====

==== Change File Name Extensions ====

Say you want to mass change file extensions in a directory.  I found this script here: [[http://www.franzone.com/2007/09/05/bash-script-to-change-file-extensions/]]

Note:  I added sort just to do it in order.

Usage:  change_extension.sh <DIRECTORY> <FROM> <TO> (and the dots get stripped so use them if you want (or not))

<code bash>
#!/bin/bash
#
OLDEXT=${2/#.}
NEWEXT=${3/#.}

find "${1}" -iname "*.${OLDEXT}" | sort |
while read F
do
  NEWFILE="${F/%${OLDEXT}/${NEWEXT}}"
  echo "mv \"${F}\" \"${NEWFILE}\""
  mv -f "${F}" "${NEWFILE}"
done

</code>

==== Lower Case Filenames ====

<code bash>
#! /bin/bash
#
# Changes every filename in working directory to all lowercase.
#
# Inspired by a script of John Dubois,
# which was translated into into Bash by Chet Ramey,
# and considerably simplified by Mendel Cooper, author of this document.


for filename in *                # Traverse all files in directory.
do
   fname=`basename $filename`
   n=`echo $fname | tr A-Z a-z`  # Change name to lowercase.
   if [ "$fname" != "$n" ]       # Rename only files not already lowercase.
   then
     mv $fname $n
   fi  
done   

exit 0
</code>

==== Path and Directory Functions ====

Here are some examples that I found [[http://www.ufoot.org/more/blog/misc/howto/bashbasename]] for yanking out the file path etc directly in bash.

<code bash>
 function basename()
 {
   local name="${1##*/}"
   echo "${name%$2}"
 }

 function dirname()
 {
   local dir="${1%${1##*/}}"
   "${dir:=./}" != "/" && dir="${dir%?}"
   echo "$dir"
 }

 # Two additional functions:
 # 1) namename prints the basename without extension
 # 2) ext prints extension of a file, including "."
 
 function namename()
 {
   local name=${1##*/}
   local name0="${name%.*}"
   echo "${name0:-$name}"
 }
 function ext()
 {
   local name=${1##*/}
   local name0="${name%.*}"
   local ext=${name0:+${name#$name0}}
   echo "${ext:-.}"
 }
</code>


==== Removing Control-M ====

Windows text files on UNIX / Linux have CONTROL-M (^M) characters all over the place.  To remove them, us **tr** (Thanks to Michelle for showing me this one):

<code bash>
tr -d '\015' <filein> > <fileout>
</code>

====== Linux ======


===== Generic =====

==== Share Network Connection ====

  * [[https://linuxhint.com/configure-nat-on-ubuntu/]] <-recent (2021)

While in the field I needed to share my wireless connection to a static IP linux machine.

  - Use UNETBootin to put the latest KDE based Linux Mint on a 2GB memory stick.
  - Boot into memory stick
  - Add wifi connection on wlan0
  - Set eth0 to static IP (192.168.0.1)
  - Configure your ethernet traffic to NAT gateway on wlan0 <code>
sudo /sbin/iptables -t nat -A POSTROUTING -o wlan0 -j MASQUERADE
# Turn on IP forwarding
echo 1 | sudo tee /proc/sys/net/ipv4/ip_forward
</code>
  - Profit!  Any computer in the 192.168.0.X subnet can now see the internet over your wifi connection.

==== udev Device Name Binding ====

  * [[https://unix.stackexchange.com/questions/66901/how-to-bind-usb-device-under-a-static-name]]


If you want a device to always come up as the same USBX device, you can set up a **udev** rule that sees the serial number of the device to assign the name.


===== CentOS =====

==== Grub Repair ====

  * see: [[http://wiki.centos.org/TipsAndTricks/ReinstallGRUB]]

===== Slackware =====

==== Window Manager ====

Selecting a default window manager is trivial, just run the following (found this [[http://www.slackware.com/config/x.php|here]]): 

  xwmconfig



===== Mythbuntu =====

==== Grabbing Schedules ====

Here is a manual way of grabbing the listings from Schedules Direct assuming you have an account.

<code>
#!/bin/bash
#
# grab the data from schedules direct
tv_grab_na_dd --dd-data $HOME/xmldata.xml

# fill the database with our listing data
mythfilldatabase --dd-file 1 -1 1 $HOME/xmldata.xml
mythbot@mythtv:~$ 
</code>

And your configuration files is: 
<code>
## using config filename $HOME/.xmltv/tv_grab_na_dd.conf
##
## our configuration file:
## which is: ~/.xmltv/tv_grab_na_dd.conf
username: USER.NAME
password: WHAT.IT.IS
timeoffset: -0400
lineup:   PC:21851
auto-config: add
channel: 16-1 WBOCDT
channel: 21-2 WBOCDT2
channel: 8 WRAVLP
channel: 47-1 WMDTDT
channel: 28-2 WCPBDT2
channel: 28-3 WCPBDT3
channel: 28-1 WCPBDT
channel: 47-2 WMDTDT2
channel: 36 WPMCCA
channel: 30 W30CI
channel: 25 W25AA
</code>

===== Ubuntu =====

==== Configuration ====

=== Determine Partition UUID ===

To get the partition UUID, there are disk dev links that will illuminate you.

<code>
ls -l /dev/disk/by-uuid
</code>

==== Grub 1 ====

=== Altering Boot Order ===

  - Modify the file **/boot/grub/menu.lst** such that **default=** the one you want to be your default

**NOTE:** each "title" entry will be counted as an item in the menu, starting with 0 as the first.

=== Fixing ===

I recently was installing [[http://www.slackware.com|Slackware]] from our laptop to a work machine, and managed to blow away the GRUB on our laptop in the process.  Here is how I fixed the GRUB boot loader, using [[https://help.ubuntu.com/community/RecoveringUbuntuAfterInstallingWindows|this page]].

  - Boot from an Ubuntu CD in Live Mode, just need a prompt.  (don't use 9.10 or above, that uses GRUB2 and will cause trouble with the older 8.04 LTS I have on this laptop)
  - Get to a prompt
  - Figure out which is our partition:  <code>
sudo fdisk -l
</code>
  - Make a mounting point for our linux partition:  <code>
sudo /media/root
</code>
  - Use **install-grub** from the CD:  (replace **/dev/sdx** with your boot drive (probably **/dev/sda**)) <code>
sudo grub-install --root-directory=/media/root /dev/sdx
</code>
  - Reboot

==== Grub 2 ====

=== Altering Boot Order ===

With **grub2** you only need to modify the file:  ****

  - Modify the contents of this file and set **DEFAULT=** to the system you want to load by default (change the time to 10 seconds also) <code>
sudo vi /etc/default/grub
</code>
  - regenerate the grub files <code>
sudo update-grub
</code>

=== Fixing ===

After upgrading from 10.04 to 11.04, grub2 got broken.  Found the answer quickly here:  [[http://www.webupd8.org/2010/05/fix-symbol-grubputs-not-found-when.html]]

==== KDE ====

=== Configure for 4K Monitor ===

High resolution monitor where icons and text are unreadable.  Need to make a couple changes to the settings to get the icons, text and toolbar to be usefuly sized!

  - Fonts
    * force fonts DPI to 130 (experiment, but 96 is too small)
  - Display and Monitor
    * Display configuration (scroll down)
    * Scale display 1.4 is what he used (again, experiment as needed)
  - Toolbar Right Click, Panel Options -> Panel settings
    * Click and hold the Height button and move around as necessary
  - Log off and back on to have them take effect

=== Replace Unity ===


Replace Unity with KDE on Ubuntu 14.04 LTS found [[http://askubuntu.com/questions/694608/how-to-install-kde-and-totally-remove-unity-on-ubuntu-14-04-lts|here]]

  - Install KDE desktop <code>
sudo apt-get install kubuntu-desktop
</code>
==== Networking ====

=== Filesystems ===

== NFS ==

You need these packages in order to mount NFS shares on your system (to act as a client)

  sudo apt-get install portmap nfs-common

Use the following to mount your drives, assuming you have the information in **/etc/fstab**

  sudo mount -a -t nfs

== NFS Server ==

In a pinch, if you want to setup a NFS server say with a boot disk, do the following:

<code>
sudo apt-get install nfs-kernel-server nfs-common portmap
sudo dpkg-reconfigure portmap
sudo /etc/init.d/portmap restart
sudo nano /etc/exports 
sudo /etc/init.d/nfs-kernel-server restart
</code>

== Samba Mounting Windows ==

To mount a windows shared directory, I found this page and took directly from it adapting it for my ReadyNAS Raid: [[http://ubuntuforums.org/showthread.php?t=280473]]

<code>
sudo mkdir /media/rabine
sudo smbmount //10.0.0.3/rabine /media/rabine -o username=MYUSERNAME,password=MYPASSWORD,uid=1000,mask=000
</code>

=== VPN ===

I got Cisco VPN working using [[http://www.craigmayhew.com/blog/2009/11/setting-up-vpn-in-ubuntu-9-10-karmic-koala/|this link]].  You just need to install the following packages for this to work:

<code>
sudo apt-get install network-manager-pptp network-manager-vpnc network-manager-openvpn
</code>

=== Wireless ===

== Drivers ==

= Broadcom BCM4312 =

This is the wireless driver for our little Dell 1501 laptop.  I guess they cannot ship the wireless firmware so you just have to run this one line, and it fixed everything.  I had to **rmmod** and just put the module back into the kernel with **modprobe** and it came up after a little pause.

<code>
sudo apt-get install b43-fwcutter
</code>

==== Packages ====

=== Must Haves ===

  * Flash Player:  <code>
sudo apt-get install flashplugin-installer
sudo apt-get install flashplugin-nonfree
</code>
  * gdl:  <code>
sudo apt-get install gnudatalanguage
</code>
  * gmt:  [[http://www.osmanoglu.org/index.php/computing/4-computing/42-highresgmtcoastlinesinubuntu]] <code>
sudo apt-get install gmt
# high resolution coast lines (get them from NOAA) and just untar these and move to /usr/share/gmt/coast/
wget ftp://ibis.grdl.noaa.gov/pub/gmt/GSHHS_high.tar.bz2
wget ftp://ibis.grdl.noaa.gov/pub/gmt/GSHHS_full.tar.bz2
</code>
  * hexedit:  <code>
sudo apt-get install hexedit
</code>
  * jed:  <code>
sudo apt-get install jed
</code>
  * md5deep:  <code>
sudo apt-get install md5deep
</code>  * SciTe Editor for Gnome (not sure about this one yet.. ) <code>
sudo apt-get install scite
</code>
  * smplayer:  <code>
sudo apt-get install smplayer
</code>
  * svn:  <code>
sudo apt-get install subversion
</code>
  * vlc:  <code>
sudo apt-get install vlc
</code>

=== Removing Old Unused Kernels ===

This removes unused kernels with a single command line:  (I found this on the third answer here: [[http://askubuntu.com/questions/590673/why-doesnt-ubuntu-remove-old-kernels-automatically]] )
<code>
sudo dpkg -l 'linux-*' | sed '/^ii/!d;/'"$(uname -r | sed "s/\(.*\)-\([^0-9]\+\)/\1/")"'/d;s/^[^ ]* [^ ]* \([^ ]*\).*/\1/;/[0-9]/!d' | xargs sudo apt-get -y purge
</code>

==== youtube-dl ====

  * This one puts it to mp3 (going to need LAME encoder for this, i think. i was using .ogg but tagging is a bitch with that) <code>
youtube-dl --extract-audio --audio-format=mp3
</code>
  * This one converts video to the target format (mkv here) <code>
youtube-dl -f mkv
</code>

====== MacOs ======

===== Scripting =====

==== bash ====

=== Files in Order ===

Sometimes I have noticed files do not happen in order when doing command line things.  Using **xargs** and **sort** allows us to sort the input files so that the order is correct.

<code>
ls -v j080_202405* | sort | xargs cat > j080_overnight.jps
</code>

====== Windoze ======

===== Cleaning Up The System =====

On the computer guys, they recommended two utilities/sites:

  - [[http://www.blackviper.com/|Black Viper]] - advice on what you can disable in your windows services.
  - [[http://www.ccleaner.com/|CC Cleaner]] - will clean up your cache and registry on your windows machines.


===== Fixing Password =====

  * [[http://pogostick.net/~pnh/ntpasswd/|NT Password]] is what you want to fix any account passwords you need.

===== Hard Disk =====

==== Drive Space ====

Where is all my drive space?  Use [[http://windirstat.info/|WinDirStat]] to find out!

==== Fixing Bootloader ====

I have a dual boot system, and I wanted to remove the GRUB loader.  Found [[http://lukeaddison.com/removing-linux-grub-restoring-windows-7-boot-gui/|this link]], and in the comments I think the correct answer is to simply run this while in Windows 7 (note.. could not find this on my drive.. so it must be on the Windows 7 install disk.. which I don't think I have!?):

<code>
bootsect /nt60 c: /mbr

</code>

===== SCP File Transfers =====

Here is a method for rate limiting your **scp** transfers in windows:

  - Download the **FileZilla Client** for windows here:  [[http://filezilla-project.org/download.php?type=client|Filezilla Client]]
  - Install and run
  - To limit your transfer speed (for example 500kbit or 50 KByte/Second) you would set things up under Edit -> Settings -> Transfers like this:  {{computers:sftpdemo2.png|}}
  - You can setup a connection under File -> Site Manager (or Control-S).  Here is my example for my local server : {{computers:sftpdemo1.png|}}
  - Then connect and setup files to transfer (let it go overnight or whatever it takes).

===== Software =====

==== VMWare ====

Say you want to run Ubuntu under Windoze.  Try this link:

  * [[http://www.howtogeek.com/howto/11287/how-to-run-ubuntu-in-windows-7-with-vmware-player/]]


====== Windoze 10 ======

===== Admin =====

==== Delete Folder Recursively ====

I wanted to remove ALL of cygwin from a windows 10 machine and was having access denied issues.  This worked:  [[https://www.adminlabs.com/how-to-remove-cygwin-permission-denied-problem]]

  - Open up a command line administrator console with
    - WindowsKey+x -> Command Prompt (Admin)
  - Take ownership of the files <code>
cd c:\
takeown /r /d y /f cygwin64
</code>
  - Give everyone full access to everything recursively <code>
icacls cygwin64 /t /grant Everyone:F
</code>
  - Blow it away!  <code>
rmdir /s /q cygwin64
</code>