If disk fails how long will it take for restoring from raid?

lowfan

I have a server with hetzner and I have made a raid1
My question is that how to know if the raid is working fine
and if some thing goes wrong how long will it take for hetzner to restore my server from the other disk? Will every thing be perfect?

LTniger

Depends. From 2h till forever.

Falzo

if it's software raid, look into mdadm and how to set up notifications.
hetzner will not be responsible for restoring your data or raid. (at least not for free)
they will just replace the disk if you find it faulty, after that you have to change your raid config and let it resync.

usually hetzner is quite quick with replacing disks, maybe a few hours. after that it depends on you, reconfiguring the software raid will take a few minutes. the resync depends on the size of the disk/array obviously.

@lowfan said: Will every thing be perfect?

if you have to ask questions like ...

@lowfan said: how to know if the raid is working fine

the answer to the former most likely is: no.

keep in mind that raid1 is no replacement for backups anyway.

tetech

How is the RAID1 configured? If software then you can check status cat /proc/mdstat. If software and a disk fails, I assume Hetzner just replace it and you take care of adding the new one to the array using mdadm.

# cat /proc/mdstat
Personalities : [raid1]
md128 : active raid1 sda1[1] sdc1[0]
      3906884608 blocks super 1.2 [2/2] [UU]
      [===============>.....]  check = 77.5% (3027932288/3906884608) finish=13744.7min speed=1065K/sec
      bitmap: 5/30 pages [20KB], 65536KB chunk

lowfan

@Falzo said:
if it's software raid, look into mdadm and how to set up notifications.
hetzner will not be responsible for restoring your data or raid. (at least not for free)
they will just replace the disk if you find it faulty, after that you have to change your raid config and let it resync.

usually hetzner is quite quick with replacing disks, maybe a few hours. after that it depends on you, reconfiguring the software raid will take a few minutes. the resync depends on the size of the disk/array obviously.

@lowfan said: Will every thing be perfect?

if you have to ask questions like ...

@lowfan said: how to know if the raid is working fine

the answer to the former most likely is: no.

keep in mind that raid1 is no replacement for backups anyway.

Your avatar is so scary I do not dare to ask more questions

lowfan

@tetech said:
How is the RAID1 configured? If software then you can check status cat /proc/mdstat. If software and a disk fails, I assume Hetzner just replace it and you take care of adding the new one to the array using mdadm.

# cat /proc/mdstat
Personalities : [raid1]
md128 : active raid1 sda1[1] sdc1[0]
      3906884608 blocks super 1.2 [2/2] [UU]
      [===============>.....]  check = 77.5% (3027932288/3906884608) finish=13744.7min speed=1065K/sec
      bitmap: 5/30 pages [20KB], 65536KB chunk

thanks If it fails hetzner will make the server to read from the healthy one ? Is there any settings I need to do?

jackb

@lowfan said:
thanks If it fails hetzner will make the server to read from the healthy one ? Is there any settings I need to do?

If a drive fails completely or exceeds the maximum number of errors specified in the mdadm configuration (usually best to leave this as default); mdadm will automatically mark the drive as failed and drop it from the array.

If you have notifications configured you should receive a notification. At that point, you can check the drive using smartctl and if it needs replaced, you should mark it as failed in any other arrays it is used in before asking hetzner to replace it. After that, you'll need to set up your partitions and add the replacement to the array.

You will need to provide at minimum the drive serial number when requesting a drive gets replaced. Most providers will also want indications that the drive is failing (i.e. SMART data / failed tests / etc)

tetech

@lowfan said:

@tetech said:
How is the RAID1 configured? If software then you can check status cat /proc/mdstat. If software and a disk fails, I assume Hetzner just replace it and you take care of adding the new one to the array using mdadm.

# cat /proc/mdstat
Personalities : [raid1]
md128 : active raid1 sda1[1] sdc1[0]
      3906884608 blocks super 1.2 [2/2] [UU]
      [===============>.....]  check = 77.5% (3027932288/3906884608) finish=13744.7min speed=1065K/sec
      bitmap: 5/30 pages [20KB], 65536KB chunk

thanks If it fails hetzner will make the server to read from the healthy one ? Is there any settings I need to do?

Yes, you would need to add the disk to the array.

@Falzo said: hetzner will not be responsible for restoring your data or raid. (at least not for free)
they will just replace the disk if you find it faulty, after that you have to change your raid config and let it resync.

WebProject

It depends on the other HDD size and age. From personal experience: on my personal PC I had very old hdd, it took more than 4h to sync data to new HDD.

lowfan

does a server management company do this for me? Or is it some thing I need to ask before I sign up?

Shazan

Yes, a management company should be able to do this for you but always better to ask before.

lowfan

@Shazan said:
Yes, a management company should be able to do this for you but always better to ask before.

OK please recommend me some good server management company that is also affordable