Parity drive upgrade

[French-Guy]

thanks for the tip..............I will make a backup once old parity drive is formatted

[French-Guy]

I'm watching videos from SpaceInvader One.....

Can i install plugins (Community Applications) when a drive is being cleared?

[Squid]

I'm watching videos from SpaceInvader One.....
Can i install plugins (Community Applications) when a drive is being cleared?

You can install anything from CA without adversely affecting the clear

Sent via telekinesis

[French-Guy]

11 hours ago, trurl said:

It is especially important after making any disk assignment change so if you need to restore flash it will boot up with the correct disk assignments.  But it is also a good idea to get a backup when you make any other change in the webUI since all webUI settings are saved to flash.

 

When it has finished clearing it will let you format the old parity as a new data disk. Wait until then to make the backup. Then the backup will know that disk is now a data disk and not parity.

Old 2TB parity drive is now cleared and formatted. And I just did a backup of the flash drive. Again, thanks for the tip

By the way, the 2TB drive was formatted in xfs, while all my other drives are reiserfs

Is this OK?

[JonathanM]

38 minutes ago, French-Guy said:

By the way, the 2TB drive was formatted in xfs, while all my other drives are reiserfs

Is this OK?

No, but not for the reason you think. Mixing format types in the array is perfectly fine. Using reiserfs is not recommended. I suggest you use the newly added XFS drive to migrate your other drives to XFS.

 

Copy all of one of your other reiserfs 2TB drives to the XFS drive, when you are sure the copy is complete and verified, change the file type of the source to XFS, and format it. Do that one at a time until all your drives are XFS.

[French-Guy]

So I should copy the content of drive #1 to the newly formatted XFS drive, check everything is fine, then format drive #1 into XFS, and copy back data onto it

And do this for all other drives

Is that correct?

How do I copy from 1 drive to another one actually?

Thanks

[JonathanM]

19 minutes ago, French-Guy said:

copy back data onto it

That part is not strictly necessary, unless you have specific disk slot numbers assigned as data destinations. Which specific drive the data is on doesn't matter for user shares.

 

Some light reading for your enjoyment. 🤣

https://wiki.unraid.net/File_System_Conversion

[French-Guy]

But i have 3 x 1,5TB and 1  x 2TB data drives....

So if I use the old parity drive (2TB - xfs) as a temporary drive to empty each drive and format them as xfs,  don't i need to coy back the content of each drive before doing the same for the next one?

 

[JonathanM]

Just use the freshly formatted XFS drive as the destination for the next reiserfs victim in line.

 

D1-2TB rfs copy to D5-2TB XFS

format D1 to XFS

D2-1.5TB rfs copy to D1-2TB XFS

format D2 to XFS

D3-1.5TB rfs copy to D2-1.5TB XFS

format D3 to XFS

D4-1.5TB rfs copy to D3-1.5TB XFS

format D4 to XFS

 

Done! End result, empty 1.5TB XFS in slot 4, all your data in slots 1,2,3,5.

 

If that arrangement bothers you, after you finish the migration you can set a new config and assign the drives with the old parity in D5 as D1 and the rest of the drives down a slot number. Since you only have 1 parity drive, parity is still valid and if you select the "parity is valid" you will be fine.

[French-Guy]

It's clear, thank you.......I don't think this new arrangement will bother me at all

One more question though: how do I copy from 1 disk to another one? Is it done thru the GUI? Do i need a specific plugin?

[JonathanM]

41 minutes ago, French-Guy said:

how do I copy from 1 disk to another one?

Either command line rsync at the local console or ssh /telnet with screen, or the unbalance plugin. There are more ways to accomplish it, but those are probably the most efficient.

[French-Guy]

OK, I will have to look at how the unBALANCE plugin works exactly....

Thanks for all your support

[French-Guy]

OK, so do I just select "from" (D4-2TB and select all the shares) and then select "to" (D5-2TB), then click on Plan?

And if there is no issue detected, I can unselect "dry run" and click on "Move"?

 

 

Edited November 21, 2018 by French-Guy

[JonathanM]

You can move if you wish, but I prefer copy as it's much faster, and you are going to be totally erasing the source drive by formatting it anyway. Just make sure there are no docker or network writes to the drives you are operating on. Reads are fine, but will slow the process a little.

[French-Guy]

ok, so let's go with a "copy" then.....I'll keep you posted 

Thanks

[French-Guy]

Ouch....................it says ~19 hrs

Speed is ~20 MB/s !

 

Is the array available anyway for streaming?

 

[JonathanM]

20 minutes ago, French-Guy said:

Is the array available anyway for streaming?

 

28 minutes ago, jonathanm said:

Reads are fine, but will slow the process a little.

 

[French-Guy]

Good morning.............................5 hours before completion 😕

You said "Copy" was much faster? I can't imagine if I had selected "Move"

Unless I have a problem somewhere???

[trurl]

45 minutes ago, French-Guy said:

Unless I have a problem somewhere???

Not obviously a problem, but no way to know without diagnostics.

[French-Guy]

i see both data drives are blinking, but also the parity drive

Does the parity drive need to be involved during the copy, or could it be "deactivated" to do the copy faster?

 

[trurl]

4 minutes ago, French-Guy said:

i see both data drives are blinking, but also the parity drive

Does the parity drive need to be involved during the copy, or could it be "deactivated" to do the copy faster?

Unraid parity is realtime and all write operations (write, delete, format are all write operations) update parity simultaneously. This is why Unraid writes slower than the speed of a simple disk write.

 

It would be possible to remove parity and then rebuild parity afterwards, but that isn't really going to buy you much because then you have to wait for parity to completely rebuild.

 

Large volumes of data just take a while even if things are working correctly. And as mentioned, without diagnostics, we can't say for sure if things are working correctly.

 

 

[French-Guy]

ok, will post a diagnostic when copy is completed......

[trurl]

5 minutes ago, French-Guy said:

ok, will post a diagnostic when copy is completed......

Diagnostic won't slow anything down. And the reason for asking for one is to see if there is anything affecting copy speed.

 

Or you could just take a look at the log in the webUI and see if you see any disk connection issues.

[French-Guy]

here it is.....

tower-diagnostics-20181122-0954.zip

[trurl]

So, looks like this is the first diagnostic you have posted to this thread. There isn't anything in the syslog about any errors but SMART for 3 of your disks are reason for concern.

 

When you look at the Error column in Main - Array Devices are they all zero?

 

When you look at the Dashboard, are you getting any warning indicator for any of your disks? I suspect you do have warnings on some disks in the Dashboard. You really should have addressed these before anything else you were doing in this thread.

 

Do you have Notifications setup to alert you immediately by email or other agent when you have any issues?

 

Relevant SMART for these disks:

 

Disk2, this attribute isn't normally monitored by Unraid, but it says FAILING NOW

Serial Number:    6XW0Y8ZN
ID# ATTRIBUTE_NAME          FLAGS    VALUE WORST THRESH FAIL RAW_VALUE
184 End-to-End_Error        -O--CK   001   001   099    NOW  112

 

Disk3, not sure if attribute 200 is important for this model or not

Serial Number:    WD-WCAVY0146918
ID# ATTRIBUTE_NAME          FLAGS    VALUE WORST THRESH FAIL RAW_VALUE
197 Current_Pending_Sector  -O--CK   200   200   000    -    128
198 Offline_Uncorrectable   ----CK   200   200   000    -    16
200 Multi_Zone_Error_Rate   ---R--   200   166   000    -    155

 

Disk 4

Serial Number:    WD-WCAZAH561359
ID# ATTRIBUTE_NAME          FLAGS    VALUE WORST THRESH FAIL RAW_VALUE
  5 Reallocated_Sector_Ct   PO--CK   200   200   140    -    13
196 Reallocated_Event_Count -O--CK   187   187   000    -    13

 

 

Not sure which should be the first to be replaced. I am guessing disk3 but I'm not sure how serious disk2 is. I wouldn't keep any of these in my array.