3Ware RAID Controller

Last change on 2023-01-30 • Created on 2020-03-19 • ID: RO-1DFAE

Introduction

This article explains how to use the 3ware RAID controller administration program. You can download it from the Hetzner mirror. (The login credentials are included in the email that you received when we at Hetzner finished your order.) You can also use the 3ware RAID controller admin program in the Rescue System.

  • When you download the program, please make sure you select the architecture suitable for your operating system (x86 for 32Bit systems and x86_64 for 64Bit systems).

  • After downloading it, extract the archive into a directory of your choice.

  • If you use it within a UNIX system (Linux/BSD), we recommend to make it executable and copy it to the /bin folder to make it convenient to use:

    cd /path/to/extracted/tw_cli/
chmod +x tw_cli
cp tw_cli /bin/tw_cli

  • You can find more information on how to use the program here on our mirror.

How to read the status of the hardware-based RAID

First, please obtain the identifier of the installed RAID controller using the following command:

tw_cli show
Ctl   Model        (V)Ports  Drives   Units   NotOpt  RRate   VRate  BBU
------------------------------------------------------------------------
c0    9750-4i      36        36       4       0       1       1      OK

In this example, the controller is identified as c0. Replace the c0 in the examples of this article with the appropriate number.

You can get the status of the RAID by using the following command:

tw_cli /c0 show
Unit  UnitType  Status         %RCmpl  %V/I/M  Stripe  Size(GB)  Cache  AVrfy
 ------------------------------------------------------------------------------
 u0    RAID-1    OK             -       -       -       698.637   ON     -

 Port   Status           Unit   Size        Blocks        Serial
 ---------------------------------------------------------------
 p0     OK               u0     698.63 GB   1465149168    S13UJ1CQ704597
 p1     OK               u0     698.63 GB   1465149168    S13UJ1BQ708871

In this example, one unit u0 is active, which consists of two drives - p0 and p1. Replace u0, p0 and p1 in the following examples with the identifiers that you obtain.

How to set up a hardware-based RAID

Regardless of whether you just need to make changes to the mode of an existing RAID, or if you need to set up a new RAID, first delete the existing unit(s):

tw_cli maint deleteunit c0 u0

As you have to specify multiple drives in the following commands, please be aware that : is used for enumerations (like comma) and - for ranges. For example p0:1:2 is the same as p0-2.

Set up RAID 0 (with two drives, not recommended!):

tw_cli maint createunit c0 rraid0 p0:1

Set up RAID 1 (with two drives):

tw_cli maint createunit c0 rraid1 p0:1

Set up RAID 5 (with three drives):

tw_cli maint createunit c0 rraid5 p0:1:2

Set up RAID 5 (with four drives):

tw_cli maint createunit c0 rraid5 p0:1:2:3

Set up RAID 5 (with four drives and a starting size of 200GB):

tw_cli
//rescue> /c0 add type=raid5 disk=0:1:2:3 v0=200

Setup RAID 6 (with four drives):

tw_cli maint createunit c0 rraid6 p0-3

How to use the drives as JBOD

The 3ware 9650SE controller can configure the drives as JBOD, which makes them usable as one large unit with the capacity of the sum of all the drives. However, LSI/3ware recommends that the drives instead be configured as single drives.

Show current state:

tw_cli /c0 show exportjbod
/c0 JBOD Export Policy = off

Enable JBOD:

tw_cli /c0 set exportjbod=on
Enabling JBOD Export Policy on /c0... Done.

How to start a REBUILD using tw_cli with a DEGRADED RAID

First, check the status of the RAID controller:

tw_cli /c0 show

Example for RAID 1:

Unit  UnitType  Status         %RCmpl  %V/I/M  Stripe  Size(GB)  Cache  AVrfy
 ------------------------------------------------------------------------------
 u0    RAID-1    DEGRADED       -       -       -       698.637   ON     -

 Port   Status           Unit   Size        Blocks        Serial
 ---------------------------------------------------------------
 p0     DEGRADED         u0     698.63 GB   1465149168    S13UJ1KS210609
 p1     OK               u0     698.63 GB   1465149168    S13UJ1NQ600102

You need to delete the DEGRADED drives from the array:

tw_cli maint remove c0 p0

If one of the drives is faulty, request a replacement drive. To do that, write a support request from on Robot. Then you need to rescan the controller:

tw_cli maint rescan c0

Now you can restart the rebuild of the array:

tw_cli maint rebuild c0 u0 p0

As soon as this is done, the rebuild will start and you can query the status:

tw_cli /c0 show rebuild

When you start the rebuild, the output will look like this:

Unit  UnitType  Status         %RCmpl  %V/I/M  Stripe  Size(GB)  Cache  AVrfy
 ------------------------------------------------------------------------------
 u0    RAID-1    REBUILDING     5       -       -       698.637   ON     -

 Port   Status           Unit   Size        Blocks        Serial
 ---------------------------------------------------------------
 p0     DEGRADED         u0     698.63 GB   1465149168    S13UJ1KS210609
 p1     OK               u0     698.63 GB   1465149168    S13UJ1NQ600102

If the rebuild is interrupted due to ECC errors, you can force a rebuild (not recommended):

tw_cli /c0/u0 start rebuild ignoreECC

Other useful commands

Show help and controller details

tw_cli help

Show controller and configuration information

tw_cli info c0

Read out the SMART values of a drive

  • For series 6000, 7000 and 8000 controllers:

    smartctl -a -d 3ware,p0 /dev/twe0

  • For series 9000 controllers:

    smartctl -a -d 3ware,p0 /dev/twa0

  • For series 9750 controllers:

    smartctl -a -d 3ware,p0 /dev/twl0

Export a drive

tw_cli /c0/p0 export

Set a drive as hot spare

tw_cli /c0 add raidtype=spare disk=2
Table of Contents