Linux

When configuring RAID it’s quite important to have the same partition tables on every disk. I’v done this many times on msdos partition tables like this: sfdisk -d /dev/sda | sfdisk /dev/sdb but it’s not working any more on GPT partition tables. Hopefully it still can be done but with different toolstack 😄 Install gdisk: apt-get install -y gdisk Then use sgdisk like this: sgdisk -R /dev/sd_dest /dev/sd_src sgdisk -G /dev/sd_dest First command will copy partition from /dev/sd_src to /dev/sd_dest....

On Debian in default installation you have different configuration files for PHP in Apache, FPM, CLI, etc. But on CentOS you have only one php.ini for all of them. In case I have, I need to have different configuration file for scripts running in CLI mode (more memory, etc). I could run it like this: php -c /etc/php-cli.ini script.php But this a little burdensome. So I do it like this:...

Everybody knows passwd command but it’s useless when you need to change ex. root password from command line without waiting for input. In such case oneliner below could help: echo "root:new_password" | chpasswd

When I was trying to update packages on one host I’ve stuck with yum hung on update. I run strace and see: strace -p 43734 Process 43734 attached - interrupt to quit futex(0x807c938, FUTEX_WAIT, 1, NULL <unfinished ...> Process 43734 detached It looks like yum database was corrupted, to repair this run: rm -f /var/lib/rpm/__db* rpm --rebuilddb yum clean all yum update Instead rm on db-files you could use gzip to have backup of these files....

I’ve few Nagios checks that require root privileges but running nrpe as root user is not acceptable. I prefer to use sudo for only these few commands. Run visudo and coment out this line: #Defaults requiretty This change is crucial to get scripts working. Then add at the end of file: %nrpe ALL=(ALL) NOPASSWD: /usr/lib64/nagios/plugins/ I’ve used nrpe group, but you have to add exactly group that your nrpe process uses....

I need to check memory usage of memcached server so I used: echo stats | nc 127.0.0.1 11211 STAT pid 2743 STAT uptime 263 STAT time 1395438951 STAT version 1.4.13 STAT pointer_size 64 STAT rusage_user 0.482926 STAT rusage_system 2.675593 STAT curr_items 8667 STAT total_items 10742 STAT bytes 23802513 STAT curr_connections 296 STAT total_connections 399 STAT connection_structures 297 STAT cmd_flush 0 STAT cmd_get 52578 STAT cmd_set 10792 STAT get_hits 28692 STAT get_misses 23886 STAT evictions 0 STAT bytes_read 35984361 STAT bytes_written 192647437 STAT limit_maxbytes 536870912 STAT threads 2 STAT accepting_conns 1 STAT listen_disabled_num 0 STAT replication MASTER STAT repcached_qi_free 8189 STAT repcached_wdata 0 STAT repcached_wsize 1026048 END For me, bytes value was important but you could find more about all statistics here  external link ....

I have development server with postfix - I wanted to allow outbound traffic to one domain but cut off all the rest. I definitely do not want that test mail or any debug info goes to service users. I have to add something like that to /etc/postfix/transport: allowed.domain.com : * discard: Then run: postmap /etc/postfix/transport At end, add these to /etc/postfix/main.cf: transport_maps = hash:/etc/postfix/transport Reload postfix: postfix reload Test if it works:...

In recent Ansible update to 1.5 version there is really nice feature ssh pipelining. This option is serious alternative to accelerated mode. Just add to you config file (ex. ~/.ansible.cfg): [ssh_connection] pipelining=True Now run any playbook - you will see the difference 😄 Source (and extended info about): http://blog.ansibleworks.com/2014/01/15/ssh-connection-upgrades-coming-in-ansible-1-5/  external link

I had quite simple task - compare two lists of hosts and check if hosts from first one are also on the second one. I started with diff: diff -u biglist.txt hosts_to_check.txt | grep -E "^\+" It was fine but output needs some filtering to get what I want. I’ve found another example with grep: grep -Fxv -f biglist.txt hosts_to_check.txt | sort -n This will search for all lines in hosts_to_check....

After reading some good opinions about MariaDB I wanted to give it a try. Upgrade looks quite straight forward but I found some issues a little tricky. Installation Add repo and key: cat > /etc/apt/sources.list <<SRC deb http://mirrors.supportex.net/mariadb/repo/5.5/debian wheezy main deb-src http://mirrors.supportex.net/mariadb/repo/5.5/debian wheezy main SRC (find more repositories here  external link ) Now install MariaDB: sudo apt-get update sudo apt-get install mariadb-server It could be better to install mariadb-server-5.5 and mariadb-client-5....