Beiträge von badiane

I forgot to mention that the modules should be loaded before you run the update-initramfs command.

This is the result of my examine command:

and for my events check:

Also as you may have noticed, your drives don't have any RAID metadata so will probably not be combined in an array since there isn't any information on how to reassemble them. Had there been metadata then the event would have shown you if the difference between data writes had the integrity of the array been altered.

Correct me if I'm wrong, but is /dev/sde your new boot disk? If from the RAID failure you didn't now shuffle the physical order of the drive then you should turn off or remove the members of the RAID array and make sure that your boot/OS disk is /dev/sda; then make the drives available. If the /dev/sda allocation doesn't persists then you should make it persistent via UDEV.

I have to go out but will be back later in the day.

I just performed a search for your hardware, so forget my question about drive locations.

Can you get me the output of mdadm --examine /dev/sd[abcde]* >> mdadm_examine_ecastellani.txt and also mdadm --examine /dev/sd[apcde]* | egrep 'Event|/dev/sd' >> mdadm_examine_event_ecastellani.txt?

Hi!

Are the additional disks directly attached to the system or are they external?

After booting, if you type lsmod | grep -raid are the modules displayed? If no, then you may need to make sure that they are built into your ramdisk. Run update-initramfs -u -k all; then reboot the machine.

Linux offers the creation of standard and non-standard RAID.

I too was interested in being to specify how I could use RAID10 in OVM and was disappointed that it forced me to use four drives. I have configured and used RAID10 with two drive sets, even when more were available. I have never had any issues with it and performances have been very good. In addition I love the fact that I don't have to worry if the drive case fails because the RAID configuration is elsewhere and as long as that is backed up, I have the ability to reconstruct the array and extract my data if need be or jump back into operation.

One thing that we have to keep in mind is that no matter what other OS' do, what matters is what Linux can and is able to do. The default response when I tell some people that I build RAID10 from two or odd numbers of drives is "... it's impossible because you need four drives for RAID10 ... " That's what they've learned and by that I mean that it's what they have been told without understanding the premise of what was inculcated and so that what it is. When, in addition to the accepted a conclusion, one attempts to further their understanding into why what they have accepted is such, they find themselves in a better position to accept change because they can track the basis or bases of the change or changes. Ideologically Linux allows things which a few modern OS' do not or even allows for their adherent to conceptualize.

I remember when I was trying to explain to someone why I was creating a LXC container inside of a VMWare virtual machine. Their argument was based on one construct of virtualization as if that was the only construct and never once stopped to question their premise and why it might be totally wrong. Linux is not only an alternative to other OS but it also offers alternative ways of accomplishing things and that requires that, though one knows what they know, they must realize that in the extent of what could be known they don't really know much. So it's the opportunities for learning and evolving that are important not how "right" we want to be or feel that we are.

There is a nice segment devoted to non-standard Linux RAID formats
https://en.wikipedia.org/wiki/Non-standard_RAID_levels

Some other pages:
http://www.ilsistemista.net/in…t-benchmark-analysis.html
https://www.suse.com/documenta…data/raidmdadmr10cpx.html
http://blog.a2o.si/2014/09/07/…aid-10-instead-of-raid-1/
I have read that Linux' RAID10 is akin to RAID1E
http://www.freeraidrecovery.com/library/raid-1e.aspx

I do hope to see interfaces which aren't interested in conforming to accepted and well known but often limited views from other OS', as opposed to ones which allow access to the potential power of the OS. The bridge, I think, lies in flexible documentation.