You may need to rename several files in a folder following a pattern.

So… you have to copy several thousand of files and folders from one server to another. And you’re short of disk space, as usual, and you want to be as efficient as possible. You can use scp -r files user@server:/... but if you do it this way you’ll notice it takes forever. This is because ssh/scp open and close a the stream for every file. So you try to zip/tar in the from side to copy just one file, and then uncompress on the target side after copy.

Sometimes we want to boot Linux with a older kernel than default. If it’s just one time, it’s a piece of cake to select it on the boot menu. Things starts to get a bit worse if you need to boot everytime with a previous version because (as it’s my case right now) you suspect there’s something that could affect performance or stability but you’re not sure. The best suited for my needs it’s telling grub2 to remember the last kernel used.

Coming from previous versions of Ubuntu, I’m getting some issues with the new configuration system on Ubuntu 18.04. Not sure if some of them are welcomed (to me) but… they are there so we have no option. Few moments ago, I’ve installed a new server with Ubuntu 18 and, thinking in future, I’ve created a blueprint on VMWare I could reuse for new servers. So I’ve create a new server with basic funcionalities, upgrade it and then convert to blueprint.

Today I realized one of my ESXi servers had system clock way long in the future. It has no internet access so I cannot simply update it with any of the multiple time servers around the world. I thought it was just a matter of simply do a date -s … but on ESXi console things are not usually so simpler. The right way to change time on console involves two commands, one for current date and another for setting the hardware clock so you get the right date again after reboot.