Adding disks to existing aggr

[scott_lau]

I am currenly running 7.2.6.1P2 and have an aggr with 16 disks and RAID size of 16. I want to increase the aggr with the 300GB FC drives available. What would be my best practice solution for this to get the best performance? Can I just add the disks to the same RAID set?

Also, when trying to expand the aggr from FileView, the disks are not listed. I have 18 spare disks available and they do show when creating a new aggr. Has anbody seen this issue?

Thanks,

Scott Lau

System Admin

HUB International Inc.

[adamfox]

I would be careful adding them to the same RAID set, although it's possible if you increase the RAID group side.

If you keep the RG size at 16 and you add disks, you will start a new RG with new parity disks (assuming RAID-DP here).  The best practice here would be add in sets of 8 or 16 for performance purposes.

As far as the FilerView issue, I'm guess I'm curious what size and type of drive are currently in your aggregate as that might affect its behavior.

[scott_lau]

The disk sizes are the same (300GB FC) listed as 268GB

[adamfox]

Ok..that's a good thing.  Assuming you have enough it should let you add them.

I'm not a FilerView expert, so you can always work with support on that part, but the CLI is pretty simple:

netapp> aggr add <aggr_name> X@300

Where X is the # of disks you want to add.  Again if you don't increase the RG, this will make your first 2 disks parity disks.

[scott_lau]

If the RG is not increased, will it create the second RG automatically (i assume...but we all know that that means right) and at what size?

[radek_kubka]

The new RG will be created automatically.

If the RG size is set to 16 & you actually add less than 16 new drives (say 11), it will create a new RG containing all new drives.

Regards,
Radek

[adamfox]

Yep.  ONTAP is smart enough to know that when you hit the RG size, it will start a new one automatically (by making the next 2 disks parity, then data drives up until the new RG hits the RG size, then it will start again.

Just be aware that adding disks to an aggregate is not easily reversible, even right after you do it.  The only way to pull a data disk out of an aggregate is to destroy the aggregate and rebuild it.  It's one of the few things in ONTAP that is difficult to undo.

[misra]

In Aggregrate Manage page, click on the aggregrate. There is a option to Add Disks to Aggregrate. And inside the wizard there is a option for manual selection. Selecting the manual option will list all the available Disks and user can select as many disk they want.

[shubhada]

As far as adding disks through FilerView is concerned,selecting "Manual" for disk selection should list all the disks.

[scott_lau]

NOPE! The list is empty even if I check the Diff RPM option!!! Some funky bug maybe in version upgrade that was done few months back! I will check with Netapp and post back later!

[danielpr]

Hi Scott,

Are you very sure all the disks belong to same disk type ? Can you please share the output of "aggr status -r"

Thanks;

Daniel

[scott_lau]

    They are the same: sorry only got half the screen! 😜

      data            5a.91         5a    5   11  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            5a.100        5a    6   4   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            5c.52         5c    3   4   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            6a.92         6a    5   12  FC:B   -  FCAL 15000 272000/557056000  274845/562884296
      data            5a.101        5a    6   5   FC:A   -  FCAL 15000 272000/557056000  274845/562884296

    RAID group /aggr12/plex0/rg1 (normal)

      RAID Disk        Device        HA  SHELF BAY CHAN Pool Type  RPM  Used (MB/blks)    Phys (MB/blks)
      ---------        ------        ------------- ---- ---- ---- ----- --------------    --------------
      dparity         6a.102        6a    6   6   FC:B   -  FCAL 15000 272000/557056000  274845/562884296
      parity          5c.53         5c    3   5   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            6a.93         6a    5   13  FC:B   -  FCAL 15000 272000/557056000  274845/562884296
      data            5a.103        5a    6   7   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            6c.54         6c    3   6   FC:B   -  FCAL 15000 272000/557056000  274845/562884296
      data            5a.104        5a    6   8   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
      data            6c.55         6c    3   7   FC:B   -  FCAL 15000 272000/557056000  274845/562884296
      data            6a.105        6a    6   9   FC:B   -  FCAL 15000 272000/557056000  274845/562884296

[danielpr]

Thought will get the spare list too .. My mistake sorry... Possible to share aggr status -s ?

Thanks;

Daniel

[scott_lau]

This is what is left after adding disks via CLI

Spare disks

RAID Disk       Device  HA  SHELF BAY CHAN Pool Type  RPM  Used (MB/blks)    Phys (MB/blks)
---------       ------  ------------- ---- ---- ---- ----- --------------    --------------
Spare disks for block or zoned checksum traditional volumes or aggregates
spare           5a.29   5a    1   13  FC:A   -  FCAL 10000 272000/557056000  280104/573653840
spare           5a.44   5a    2   12  FC:A   -  FCAL 10000 272000/557056000  280104/573653840
spare           6a.43   6a    2   11  FC:B   -  FCAL 10000 272000/557056000  280104/573653840
spare           5a.107  5a    6   11  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5a.108  5a    6   12  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5c.56   5c    3   8   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5c.57   5c    3   9   FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5c.58   5c    3   10  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5c.59   5c    3   11  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           5c.61   5c    3   13  FC:A   -  FCAL 15000 272000/557056000  274845/562884296
spare           6a.106  6a    6   10  FC:B   -  FCAL 15000 272000/557056000  274845/562884296
spare           6a.109  6a    6   13  FC:B   -  FCAL 15000 272000/557056000  274845/562884296
spare           6c.60   6c    3   12  FC:B   -  FCAL 15000 272000/557056000  274845/562884296

[AVANISHRANJAN]

Hey Scott..

Why you want to increase the Raid Size it a little bit hamper performance. 16 is the right Raid size.

make a new raid group only by command agg add <aggr name> -d <disk list> ... <disk list>

please note that as i can see you have different type of RPM disks so it can cause performance issues so add the spares of same RPM type.