Difference between revisions of "Talk:RAID"

From ArchWiki
Jump to: navigation, search
(Major re-write)
m (GPT partitions: fix a word)
 
(21 intermediate revisions by 8 users not shown)
Line 14: Line 14:
  
 
:: This is here because it's how I prepare hard-drives before setting them up for RAID.  Not everyone uses GPT *yet* so didn't want to just stick it on the main page..  ~ [[User:AskApache|AskApache]] ([[User talk:AskApache|talk]]) 09:26, 3 October 2013 (UTC)
 
:: This is here because it's how I prepare hard-drives before setting them up for RAID.  Not everyone uses GPT *yet* so didn't want to just stick it on the main page..  ~ [[User:AskApache|AskApache]] ([[User talk:AskApache|talk]]) 09:26, 3 October 2013 (UTC)
 +
 +
::: This is nice, especially if you want to script partitioning in a more readable way than piping input into fdisk or gdisk. --[[User:Nearwood|Nearwood]] ([[User talk:Nearwood|talk]]) 18:16, 15 March 2017 (UTC)
  
 
== Major re-write ==
 
== Major re-write ==
 
I've done a pretty major overhaul to the article over the past week.  Please check it for accuracy.  One of my goals to was add a thread of continuity to the article so it reads as complete work rather than as a hodgepodge of advice.  I feel that mixing formatting types and utils for example is confusing to newbies.  I recommend sticking with GTP as you can see in the text.[[User:Graysky|Graysky]] ([[User talk:Graysky|talk]]) 23:22, 5 October 2013 (UTC)
 
I've done a pretty major overhaul to the article over the past week.  Please check it for accuracy.  One of my goals to was add a thread of continuity to the article so it reads as complete work rather than as a hodgepodge of advice.  I feel that mixing formatting types and utils for example is confusing to newbies.  I recommend sticking with GTP as you can see in the text.[[User:Graysky|Graysky]] ([[User talk:Graysky|talk]]) 23:22, 5 October 2013 (UTC)
 +
 +
== RAID 1 and Stride/Stripe ==
 +
 +
The section "Build the Array" mentions "In a RAID1 the chunk switch is actually not needed." and mdadm outputs "chunk size ignored for this level". cat /proc/mdstat outputs "65536KB chunk" regardless of what chunk size was chosen during creating.
 +
 +
Yet the section "Calculating the Stride and Stripe-width" has an example for RAID1 and uses a 64KB chunk size for calculating it. What is this math based on if it is impossible to choose a chunk size for RAID1
 +
{{unsigned|13 January 2014|Malstrond}}
 +
 +
:: i think you are right. the man page says "[chunk] is  only  meaningful  for RAID0, RAID4, RAID5, RAID6, and RAID10.". raid 1 is a single chunck. so it ignores that flag. ~ [[User:Gcb|Gcb]] ([[User talk:Gcb|talk]])
 +
 +
----
 +
 +
:: Note that {{ic|/proc/mdstat}} has two "chunks" listed:
 +
  Personalities : [raid6] [raid5] [raid4]
 +
  md0 : active raid6 sda2[0] sdb2[1] sde2[4] sdd2[3] sdc2[2]
 +
      2929501200 blocks super 1.2 level 6, '''16k chunk''', algorithm 2 [5/5] [UUUUU]
 +
      bitmap: 1/8 pages [4KB], '''65536KB chunk'''
 +
 +
:: The former is my specified chunk, the latter is for some other specification unknown to me.
 +
:: --[[User:Nearwood|Nearwood]] ([[User talk:Nearwood|talk]]) 18:13, 15 March 2017 (UTC)
 +
 +
== Subsituting one identical disk from another ==
 +
 +
It's useful to use sfdisk -d /dev/sda | sfdisk /dev/sdb for copy the partition from one of the disks of the raid to the replacing disk. External references: http://www.howtoforge.com/replacing_hard_disks_in_a_raid1_array. If you don't use sfdisk, then you could receive the error: mdadm: /dev/sdb1 not large enough to join array
 +
{{unsigned|8 November 2014|Xan}}
 +
 +
== <s>"possible issues"</s> ==
 +
 +
> To prevent possible issues each device in the RAID should be securely wiped.
 +
 +
What are the issues exactly? What can happen if the device isn't wiped before use?
 +
{{unsigned|01:18, 25 June 2016‎|Nowaker}}
 +
 +
:The issue is more that it is very hard to wipe when it is running, so more a pointer. I've reworded the sentence.[https://wiki.archlinux.org/index.php?title=RAID&type=revision&diff=459919&oldid=459153] Closing. --[[User:Indigo|Indigo]] ([[User talk:Indigo|talk]]) 10:42, 23 December 2016 (UTC)

Latest revision as of 18:17, 15 March 2017

GPT partitions

zap (destroy) GPT and MBR data structures

 sgdisk --zap-all /dev/sdb

create largest possible new partition

 sgdisk --largest-new=1 /dev/sdb

check partition table integrity

 sgdisk --verify /dev/sdb

print partition table

 sgdisk --print /dev/sdb
Is this a mis-paste? I can't quite see why it is here?
jasonwryan (talk) 00:36, 19 July 2013 (UTC)
This is here because it's how I prepare hard-drives before setting them up for RAID. Not everyone uses GPT *yet* so didn't want to just stick it on the main page.. ~ AskApache (talk) 09:26, 3 October 2013 (UTC)
This is nice, especially if you want to script partitioning in a more readable way than piping input into fdisk or gdisk. --Nearwood (talk) 18:16, 15 March 2017 (UTC)

Major re-write

I've done a pretty major overhaul to the article over the past week. Please check it for accuracy. One of my goals to was add a thread of continuity to the article so it reads as complete work rather than as a hodgepodge of advice. I feel that mixing formatting types and utils for example is confusing to newbies. I recommend sticking with GTP as you can see in the text.Graysky (talk) 23:22, 5 October 2013 (UTC)

RAID 1 and Stride/Stripe

The section "Build the Array" mentions "In a RAID1 the chunk switch is actually not needed." and mdadm outputs "chunk size ignored for this level". cat /proc/mdstat outputs "65536KB chunk" regardless of what chunk size was chosen during creating.

Yet the section "Calculating the Stride and Stripe-width" has an example for RAID1 and uses a 64KB chunk size for calculating it. What is this math based on if it is impossible to choose a chunk size for RAID1 —This unsigned comment is by Malstrond (talk) 13 January 2014. Please sign your posts with ~~~~!

i think you are right. the man page says "[chunk] is only meaningful for RAID0, RAID4, RAID5, RAID6, and RAID10.". raid 1 is a single chunck. so it ignores that flag. ~ Gcb (talk)

Note that /proc/mdstat has two "chunks" listed:
 Personalities : [raid6] [raid5] [raid4]
 md0 : active raid6 sda2[0] sdb2[1] sde2[4] sdd2[3] sdc2[2]
     2929501200 blocks super 1.2 level 6, 16k chunk, algorithm 2 [5/5] [UUUUU]
     bitmap: 1/8 pages [4KB], 65536KB chunk
The former is my specified chunk, the latter is for some other specification unknown to me.
--Nearwood (talk) 18:13, 15 March 2017 (UTC)

Subsituting one identical disk from another

It's useful to use sfdisk -d /dev/sda | sfdisk /dev/sdb for copy the partition from one of the disks of the raid to the replacing disk. External references: http://www.howtoforge.com/replacing_hard_disks_in_a_raid1_array. If you don't use sfdisk, then you could receive the error: mdadm: /dev/sdb1 not large enough to join array —This unsigned comment is by Xan (talk) 8 November 2014. Please sign your posts with ~~~~!

"possible issues"

> To prevent possible issues each device in the RAID should be securely wiped.

What are the issues exactly? What can happen if the device isn't wiped before use? —This unsigned comment is by Nowaker (talk) 01:18, 25 June 2016‎. Please sign your posts with ~~~~!

The issue is more that it is very hard to wipe when it is running, so more a pointer. I've reworded the sentence.[1] Closing. --Indigo (talk) 10:42, 23 December 2016 (UTC)