Migration from RAID1 the RAID5 turned out to be much easier than I thought 🙄 are broadly 5 simple steps a little waiting and 1 a beer for courage.
To me the system has created RAID array md0 involving 2 disk sda and sdb. I'll add to them the 3rd sdc to create the RAID5 3 disc. Generally this is for scientific purposes, the acrobatics of virtualka I have not yet tested in real environment, but don't expect dramas and the real time machine.
Create the same file system layout as other drives us – sfdisk -d /dev/sdb | sfdisk /dev/sdc
Build upon our current array of RAID5 – mdadm –grow /dev/md0 –level=5
Add the new disk to the array – mdadm –manage /dev/md0 –add /dev/sdc . Here comes the thin point that the array is still going to be RAID1 and won't start sync because the new disc is spare
The most important moment sdc becomes active and starts the synchronization – mdadm –grow /dev/md0 –raid-devices=3 . A good time to open the beer if not done 😉 stay in no case process!!!
Once completed the synchronization stuck to resize the partition because the loss of space in 1/e RAID1 and RAID5 for n e 1-1/n
The big bonus is that you don't have a system restart or pulling and making additional arrays.
Today I think the ponder on this perversion of nature CentOS. Inspired by the recently published 6th CentOS version I have and what I mean. In itself, this shit is development RedHat and it is on their server fork Red Had Linux Enterprise. Use the rpm package manager them (which is terribly great, Yes, but not).
Let me start by saying that made me begin to write and to ponder what kind of freak this is CentOS and soil in my servers. About a week ago came out verisâ 6 and I decided to update my current 5.6 installation of our VPS hosting. I was very unpleasantly surprised when I saw the, I don't consider the update packages. I decided, that something I dabble and checked out on the Internet. I was shocked when I saw the recommendation is to do a clean install and upgrade from distributiviven 5.6 is not recommended and is made with pounds of black magic and that is why it is not possible in the standard way. Hmmm pretty interesting point. 😆 and that's enterprise distribution, very interesting. I don't see how you can even be classified in this category except, that manufacturer and put this resounding name. Assume the following 2 script – one of them is doing a new install the other is not doing.
1. Scenarios – I'm downloading it razkačam server stops all services that support. Install it for 1 hour or more business for which I work is undergoing major losses as money. I work as a system will probably take adminsnitrator or salty fines. Comment not all sorrows along the correct settings and the archive data and settings. The keel is rezulatta nerves have a clean system. Obvious options is not acceptable.
2. Scenarios – We don't do a distribution upgrade system sits so while running the security patches. While at one point and not terminate maintenance after a while gets hacked because of a hole in any of the services offered, for nev″zmožnosta to ensure distrubutivno upgrade. Steal data or simply just to compromise server – salty back fines or you lose your job.
Quite interesting and both the script end quite unpleasant for their system administrator for capital mistake in the design of selection and distribution company that laziness to not ensure compatibility between packages. While on the other hand there is not so much enterprise distributions which are quietly, without big names. I have a Debian Server that is version 3 upgraded to version aktualanta 6 now that he's been through 3 major upgrades like there was no refusal to access services. In principle, one of the main principle of the admin's – “If you work alone” but that's why people discover holes placed patches so they come out new packages to improve stability or to accelerate productivity. In conclusion, in addition to my personal opinion but the opinion of many of my friends where the good Admin than me is CenOS sucks.