Due to my busy work schedule my personal web site was dated. So much so that it was a few releases behind schedule. I had built it using Ghost (headless Node.js CMS) and by quickly building some Docker containers I managed to upgrade the database relatively fast. I have built many data driven dynamic sites most based on PHP or Perl. However, it was also time to re-evaluate whether to continue with Ghost.

These are index folders, the presence of which can be quite annoying. To locate them find . -type d -name "@eaDir" if you’re feeling ok about automatically deleting them then: find . -type d -name "@eaDir" -print0 | xargs -0 rm -rf On a Synology NAS you can disable feature as follows synoservice --disable pkgctl-SynoFinder warn: The runkey of service [pkgctl-SynoFinder] has been set to “No”

I use nfs quite a lot. It allows local access to remote files i.e. the remote folder is mounted locally so you can access the remote files locally. Often overlooked, make sure the proper NFS RPC-based services are enabled for portmap (not required for nfs4). Issue the following command as root: rpcinfo -p This should list some lines mentioning mountd. To install nfs (here on centos): yum install nfs-utils nfs-utils-lib Proper install details are descriped in https://www.

The agility that Open Source produces is tremendous but only as effective as the skills of your top IT person. The learning curve is steep but tremendously rewarding. I have collected here a repository of code and other material that I have struggled with at some stage or the other. I hope it is of use. Enjoy and do spread the word of Linux.

If you run a cron job and you would like to only see the errors reported use cronic on Centos or chrronic on Ubuntu. The latter comes with the package moreutils apt-get install moreutils

At times you see that a particular image or so is not found when loading a web page and you are at a loss where the call came from. At least I had that today when I noticed that a menu image was called. The command to search for that is actually really straight forward: grep -rnw 'directory' -e "string pattern" e.g. grep -rnw ‘public/’ -e mobile-menu.png

I use yui-compressor to minify my files. It’s so easy. In Ubuntu (15.04) it’s available via apt-get. To install it (on Ubuntu) type the following in your terminal window apt-get install yui-compressor The command structure to run it is yui-compressor SOURCE > TARGET e.g. yui-compressor public/css/custom.css > public/css/custom.min.css

A somewhat weird but simple way to reduce a pdf is to concert it to a ps and back e.g. pdf2ps input.pdf output.ps && ps2pdf output.ps output.pdf

I needed to generate a .htaccess redirect file where each line has a simple command redirect 301 old-file new-file The file called crawl-error.txt with all the required urls that needed a redirect was quite long so I did the following in no time. First ensure that the file is really a proper text file as otherwise whatever you want to append will be prepended instead. dos2unix crawl-error.txt Then prepend the text, in my case “redirect 301 “, and output the result to .

I discovered that it is not possible, at least not for the same user, to have two git projects in the same account as github uses the ssh key to identify your repository. Once I realized that this was the problem, the quick fix was to clone the repository in another account. I created a new user and set it up from there. I created a ssh key which I copied across onto the github site.

I have a web server which needs to link to a MS SQL server (not my choice). In my case I use Plack which calls the database in four steps. The catch I often struggle with is that some other files with the same name may float around and which you may not be aware off. Run therefore a “find / -name odbc.ini” to double check this. Ditto for the others.

I did a fresh install the other day of Ubuntu 14.04 on a little notebook for my daughter. All worked out of the box with the exception of the wireless which had worked before using version 12. The following terminal commands may give you a clue as to which network controller you have. iwconfig lspci -vnn | grep Network All I had to do was install a driver and reboot

No doubt you have, like me, one or more documents that you would like extra protected. How can you remember all your passwords, access codes, urls etc which you know you should not have available in plain text. In Linux there are several tools that make file encrypting easy. I like to use a little tool called ccrypt for its simplicity. Having said that, it uses an algorithm known as AES, with 256-bit considered an extremely powerful encryption method.

After my server was attacked and brought down by the Chinese I wasn’t certain whether or not the machine was compromised. I decided though to move my mail to Google Apps instead, preferring a good sleep over worries of another more ferocious attack. As my mail wasn’t on a live machine I could not use Google’s suggestion of using their API. Copying mail using both accounts in Thunderbird would have been a great manual effort but I also noticed it to be highly unreliable for large folders so I swiftly gave that up.

After the Chinese attacked my system trice recently I am a bit paranoid about securing my system. I used to sync data using rsync which uses ssh. However, using this as root is not a good idea hence I decided to use rsync as a non-root user without a prompt for a password. This post is therefore an update from the one I wrote almost exactly two years ago. First the essentials

You use a Postfix setup with amavisd-new and get the following errors in your mail log: status=deferred (connect to 127.0.0.1[127.0.0.1]:10024 Connection refused) The solution is to restart both amavisd-new and Postfix and then check with netstat -tap that both are running on the ports 25 and 10025 (postfix) and 10024 (amavisd-new).

If you get many of these in your massages log, it is as a result of spammers hitting the mail server with bogus domains and the DNS is trying to resolve the IP to the domain, it is coming back false, and the mail server drops the message as designed. To avoid these message, though I like knowing what happens, add the following in the logging section of named.conf category lame-servers {null;};

I noticed a strange error while uploading files. My server does not allow any file to be uploaded above the 128kb, even when the settings in /etc/php.ini are set to higher sizes. After a lot of searching I managed to find the solution: add the following to the file /etc/httpd/conf.d/fcgid.conf : FcgidMaxRequestLen 1073741824 Don’t forget to restart the apache server.

HTML development work often is done remotely with all dangers of messing up a production version. My friend Gurunandan Bhat suggested I set up a local development environment. As always all credit goes to him for his support. This is therefore a howto for anyone eager to set up a local development environment for a dynamic web site. For that we use Perlbrew, Dancer, Plack, Template Toolkit, Memcached. Perlbrew is a tool to manage multiple perl installations in your $HOME directory.

Set up a git account on github.com. Upload your ssh key. To verify this is done click on top right Account Settings, then ssh keys on left. Then proceed to create a repository on Github. Don’t add a README. On your local machine, go into your application folder. Type git init This creates a .git folder inside the application folder. git commit -m "Initial Commit" This will come back with nothing added to commit but untracked files present (use “git add” to track).

I have been receiving alerts from root that “Device: /dev/sda [SAT], 4 Currently unreadable (pending) sectors”. As all my data is backup properly I don’t worry about it. Googling it, I didn’t find a straight answer for this. smartctl –all /dev/sda | grep -e “Reallocated_Sector_Ct” -e “Current_Pending_Sector” -e “Offline_Uncorrectable” -e “UDMA_CRC_Error_Count” -e “Hardware_ECC_Recovered” The output is as follows: [root@myhome ~]# smartctl –all /dev/sda | grep -e “Reallocated_Sector_Ct” -e “Current_Pending_Sector” -e “Offline_Uncorrectable” -e “UDMA_CRC_Error_Count” -e “Hardware_ECC_Recovered” 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 4 198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 0

Normally your photos in your digital camera are saved with some number which says very little and may conflict with the next batch after you clear your memory. One of the best programs I have ever used to rename my photos is jhead. This takes the date when the picture was actually taken. Fabulous! Try it out, I promise you won’t regret it. jhead -n%Y%m%d-%H%M%S * One improvement would be to add the gps coordinates but my camera is not that fancy as yet.

Want to know when your server rebooted lately? Simple last reboot or in case you really only want to see the very last one do who -b

As I often work on multiple servers it is rather dangerous that one give a command on the wrong server. To minimise the risk it is good practise to colour the terminal prompt by adding this at the end of the ~/.bash_profile file. export PS1="\e[0;32m\u\e[m\e[0;32m@\h\w \$ \e[m" and something like this in the root of the same server (i.e. you highlight root red vs home green) export PS1="\e[0;31m\u\e[m\e[0;32m@\h\w \$ \e[m" The numericals for the various colours you can use are:

I was looking at creating a simple alias to enable to send a mail out to everybody. My first thought was to set an alias in /etc/aliases but that didn’t work as the email is was not recognized by postfix. I use ISPconfig, the latest version of which is supposed to support mailman. Hence I thought to quickly install that but ran into some minor problems. The main problem being that Google shows very few posts related to my problems.

When an employee leaves you may want to pass on the ownership of his/her files to someone else. To do that simply give the following command: chown -cR --from:olduser newuser:newgroup *

A staff member came to me to state that his Yahoo account had been hacked from our company account, or so he claimed. This made me look at the samba logs and realise that user logging was not actually implemented while it is relatively easy to do so. This is used on Ubuntu 11.04 but the same should more or less work on any other system. Edit the Samba configuration file and add the following

systemctl enable name-of-the-service To use Samba as an example systemctl enable smb.service systemctl enable nmb.service systemctl start smb.service systemctl start nmb.service All that the first command does is create a symlink ln -s ‘/usr/lib/systemd/system/smb.service’ ‘/etc/systemd/system/multi-user.target.wants/smb.service’ Thus /etc/systemd/system/multi-user.target.wants contains all services that need to start automatically.

To make a tape backup, give the following two commands on the terminal of the Felicity Linux server. The first one rewinds the tape to the beginning and the second one makes the actual backup. Just cut and paste the following two lines, press enter, and leave it for the machine to do the backup. mt -f /dev/st0 rewind tar -cpzvf /dev/st0 /home/ So only the /home/ directory is being backed up.

I have a Sony home theatre and I have been bugged for ages as only some of my video files where showing up. A real bummer. After lots of searching and trying different programs, which all had some issue, I bumped into handbrake. Before you try and install handbrake on Fedora 17 first install the following packages yum groupinstall "Development Tools" "Development Libraries" "X Software Development" "GNOME Software Development" yasm zlib-devel bzip2-devel fribidi-devel dbus-glib-devel libgudev1-devel webkitgtk-devel libnotify-devel gstreamer-devel gstreamer-plugins-base-devel

Apply sa-learn to all email boxes called Junk in a cron job /usr/bin/sa-learn --spam /var/vmail/*/*/.Junk/*/* /bin/rm -rf /var/vmail/*/*/.Junk/cur/*

Use Ghostscript to contatenate pdf files: gs -dBATCH -dNOPAUSE -q -sDEVICE=pdfwrite -sOutputFile=finished.pdf file1.pdf file2.pdf

For some reason my postfix installation went totally wrong. I don’t know why other that I suspect that my IT manager may have given the wrong command and in his enthusiasm messed things up. I didn’t see much of an option but to reinstall postfix but that turned to be a lot harder than expected as aptitude insisted to install exim as a replacement which it couldn’t do. No matter what purge postfix etc I tried, I was stuck.

Uploading a file is easy if you see the upload link. I didn’t. To ensure you see it you have to change the setting in LocalSettings.php $wgEnableUploads = true; You may get an error **Could not create directory “mwstore://local-backend/local-public/ **and some more alpha characters. This turned out a simple permission problem. You have to set the permission of your images directory to 777 chmod 777

To find out whether the pc you are working on is actually a 32 or 64 bit machine you have to give the following terminal command: uname -m If it displays i386 or i686 then it is 32 bit. x86_64 is a 64 bit machine.

Create a small bash file vi ~/thunder and cut and paste the following in it. #!/bin/sh export LC_TIME=en_GB exec /usr/bin/thunderbird "$@ Then make the file executable and edit the launcher chmod 774 ~/thunder gnome-desktop-item-edit /usr/share/applications/mozilla-thunderbird.desktop and replace the command with ~/thunder i.e. the path of the above file. Note: unfortunately you will have to repeat the last command after each upgrade

I was surprised that apache didn’t work out of the box. Thanks to DoomDeathMetal (what a name!) this was easily solved: rm -rf /etc/httpd yum -y reinstall httpd systemctl start httpd.service

I use rsync for all my backup scripts. It is fast and easy. #!/bin/bash LOGFILE=/var/log/backup.log EMAILID=Alfred Tuinman@xyz.com NASDISK=192.168.x.x:/volume1/backup/myserver # the NAS station ARGMNTS="-rptgoDLKve ssh --delete-excluded --safe-links --exclude-from=/path-to-my-file/rsync_exclude" ERROR=0 log='' # for a daily copy on weekdays (5 copies) if [ `date +%u` -lt 6 ]; then weekday=`date +%w-%A` else weekday='5-Friday'; fi # for rotating every day (2 copies) uncomment the following line in lieu of the above #if [ $(( `date +%w` % 2 )) -eq 0 ]; then weekday=1 else weekday=2 fi command="/var/vmail/ $NASDISK/$weekday/var/vmail/" echo "syncing $command" if rsync $ARGMNTS $command then log="$log `date +%d-%m-%G\ %k:%M:%m` | `basename $0` | Synchronised | $command " else ERROR=1 log="$log `date +%d-%m-%G\ %k:%M:%m` | `basename $0` | Error | $command ($?

You have to add the repository to be able to get plugins for mp3 yum localinstall --nogpgcheck http://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-stable.noarch.rpm http://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-stable.noarch.rpm Then ad the actual plugin yum install gstreamer-plugins-ugly gstreamer-plugins-bad-nonfree

Sometimes Java isn’t installed by default. To add it sudo apt-get install icedtea6-plugin

I know that you don’t need a GUI op a server but sometimes you are in a small business environment where the people having to use it are not that familiar with a terminal. If you install ubuntu-desktop you overkill it by installing all user software packages like office, some games,gimp, etc which of course will also take ages. The following will set us a basic gui quickly. sudo aptitude install --without-recommends ubuntu-desktop

I did try to in vain install vmware so I changed to virtualbox. There I got the following error: “Windows failed to start. A recent hardware or software change might be the cause. To fix the problem: 1. Insert your windows install disk and restart the computer. 2. Choose your language settings, and then click NEXT 3. Click repair the computer. To avoid this, click on Settings tab of the Virtualbox.

I was struggling to understand why cron, which was running ok, was not executing my file in /etc/cron.hourly/hourly.backup while if I would run this script by hand it was working file! The reason is that cron doesn’t like a dot in a filename (at least in Ubuntu)

It is handy to have cron log to /var/log/cron.log. You can change this setting in sudo vim /etc/rsyslog.d/50-default.conf You will need to restart the services sudo service rsyslog restart sudo service cron restart

I always set the permissions for the main shared folders in samba to 3770. This way any files made below this will inherit the group ownership of the parent.

I always thought this was such a handy utility. In Ubuntu there is something similar to check if services are set to start at boot up. apt-get install sysv-rc-conf

in mediawiki change the url suffix to http://my-domain.com/my-wiki/index.php?title=MediaWiki:Sidebar&action=edit

This error happened to the mysql user and password not being correct on file. Change the setting in LocalSettings.php. You can verify whether the details you have are correct by logging into the mysql database mysql -u ''wikiuser'' -p ''wikipassword'' where of course you have to replace the user name and password with your own details.

debian-sys-maint is a system MySQL user created to be able to start/stop the databases and to carry out maintenance operations. The above error message is therefore due to a mismatch between the mysql debian-sys-maint password in the mysql database and that in the configuration file /etc/debian.cnf. Check and copy the password on file: cat /etc/debian.cnf log into mysql mysql -u root -p set the mysql database password for debian-sys-maint to that as on file in the /etc/debian.

unable to start piped log program ’ /usr/local/ispconfig/server/scripts/vlogger -s access.log -t “%Y%m%d-access.log” -d “/etc/vlogger-dbi.conf” /var/log/ispconfig/httpd’: No such file or directory Unable to open logs This happened after I deleted ISPConfig3 which for unknown reasons I could not get to work out of the box. I checked the contents off all the other apache files and bingo I found the culprit. cat /etc/apache2/sites-enabled/000-ispconfig.conf |grep vlogger ISPConfig Logfile configuration for vlogger CustomLog “| /usr/local/ispconfig/server/scripts/vlogger -s access.

The SynologyAssistant had worked out of the box on Ubuntu but for FC17 I had to do some extra work. First of all I had to add quite a few packages: yum install libXt.so.6 libXext.so.6 libfontconfig.so.1 libXrender.so.1 libSM.so.6 libgobject-2.0.so.0 libfreetype.so.6 libstdc* glib glib-devel glib2 libqwt In the firewall I added eth0 as a trusted device. FC17 uses bios device names like p17p1 rather than NIC names as eth0. Not sure whether that matters but I changed it to the old nic name by appending ”biosdevname=0” in the command line starting with the word kernel in grub.

I use xsane which needs to be added as well as some other development packages yum install xsane gcc cups-devel python-devel libusb-devel libtool libjpeg-turbo-devel qt4 qt4-devel pyqt4 pyqt4-devel dbus-devel Run hp-check to see if you have all the required packages covered hp-check I still haven’t got my scanner to work

For my Adobe flash plugin to work I had to install a few extra packages yum install flash-plugin nspluginwrapper alsa-plugins-pulseaudio libcurl also the plugins for mp3 etc gstreamer-plugins-bad-nonfree

While it may be nice to see a clean desktop I do find it handy to keep my current work load there. To show that again follow these menu options Activities -> Applications -> Advanced Settings -> Desktop -> Have file manager handle the desktop

Activities -> Applications -> Advanced Settings -> Shell -> Show date in clock

To backup a single database: mysqldump --user=root --password=password mydb > mydb.sql Yes it is no typo, there is no space between the -p and the password! You may like to add this to the my.cfg file in etc to avoid [mysqldump] events ignore-table=mysql.events If you didn’t use the –all-databases option but instead selectively backed up one or more tables or databases, you need to tell MySQL which database to place them in when restoring them.

find . -type f -print0 | xargs -0 chmod 444 Change the type from f to d if you only want to change the permissions on the directories, Plus the permission of course. find . -type d -print0 | xargs -0 chmod 755

find . -type f -name ‘[[:upper:]]*’ If you want to change files to all lower case, e.g. your mp3 music files then run the following script. The -i option for the mv command stands for interactive, i.e. it will prompt for your approval. #!/bin/sh if [ $# -eq 0 ] ; then echo Usage: $0 Files exit 0 fi for f in $* ; do g:`echo $f | tr "[A-Z]" "[a-z]"` echo mv -i $f $g mv -i $f $g done Thanks to Anil Awasare for this script

Download the new mediawiki from their website to the html directory wget http://download.wikimedia.org/mediawiki/1.17/mediawiki-1.17.0.tar.gz Extract the download and type tar -zxvf mediawiki-1.17.0.tar.gz to extract the file to the current directory. Rename the extracted directory to something easier mv mediawiki-1.17.0 update Copy the images directory and LocalSettings.php from the old to the new wiki rsync old_directory/images/ update/images/ cp old_directory/LocalSettings.php update/LocalSettings.php Backup the old directory tar -zcvf archive-name.tar.gz directory-name Remove the old directory

If you get this irritating pop-up all the time, go to programs and type “Password” a program utility called “Password and keys” will be displayed. Right click on the displayed item called “Login”. This will ask for the old and new password. If you have not changed your user password, just enter the current password for both old and new.

Recently I installed Ubuntu 12.04 server only to find that the samba server was horribly slow. I replaced it with Fedora 15 and it runs like a train. Yes, I know 15 is an old version but for a production site you don’t want the latest bugs. Also, howtoforge has a stunningly good article on how to set up a perfect server. Minimize and re-size buttons So I have been working with older Fedora desktop versions before but FC17 threw me a bit.

Use the sed terminal command with an input file called say test.html sed 's/<[^>]*>//g' input.html if you want to save the output just do the following sed 's/<[^>]*>//g' input.html > output.txt