D
D
DmitryKoterov2013-04-13 02:38:39
linux
DmitryKoterov, 2013-04-13 02:38:39

What is the reliability of software RAID-1?

I here thought that RAID1 rescues from something. Well, or at least saves when one of the disks dies. Okay, at least when the SECOND disk dies (if the first one is bootable).

But that doesn't seem to be the case, does it? Here, for example, is just the case - the second disk died (it just disappeared from the system suddenly, as if it had never existed), they replaced it with another one in the data center, the machine was rebooted ... but it does not rise. I was not too lazy to ask KVM, and saw there that the boot process was looping (in the screenshot below, the second screen, which is below, repeats over and over again with an interval of 20 seconds):

Screenshot: www.dropbox.com/s/kviqaiv234225sf/raid.jpg

, why is that? Isn't RAID supposed to protect against this kind of failure?

And then I just went into recovery mode, partitioned the second disk with sgdisk, added it to RAID, rebooted - and the system got up. Why didn't she want to climb on one disc?

Answer the question

In order to leave comments, you need to log in

2 answer(s)
L
lesha_penguin, 2013-04-13
@lesha_penguin

You have "misunderstandings" due to the fact that you mixed two completely different concepts "raid1 reliability" and "distribution features".
As for the first point (raid1 reliability) - everything is just fine with you! ;) What is the main task of raid1? Save information! And it looks like he can handle it! Yes, your disk crashed, they replaced it with a new one - the raid was rebuilt - the information is intact - it means everything is OK! Raid1 completed his task! Just as planned!
As for the second point, which actually caused your question. A specific distribution, at the time of loading, seeing one or more degraded raid arrays, did not want to start “in normal mode”. Well, these are the features of a particular distribution.

S
serjs, 2013-04-14
@serjs

In deb systems, it seems like after installing the system on a soft raid from the OS, after installing it, put the bootloader on each disk different from the main bootloader, if your 2nd disk crashed, most likely it was bootable, when the disk crashed from the array, the boot does not break ideologically.
PS The screenshot is not available, I don't know what's there, but can you /boot not be a raid? Or he has his own separate raid (they like to do this not infrequently), but again, the download comes from a specific disk

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question