cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
721
Views
0
Helpful
4
Replies

Allocating RAID groups to local disks on Cisco UCS C240 M4SX

JDMils
Level 1
Level 1

I'm getting no where trying to get this server up and running because I can't find a way to allocate the RAID groups to the locally installed SSD disks.

The server has the following disks in the following order:

DiskSize (GB)
1960
2960
3960
4960
5960
6960
7960
8960
9300
10300
11240
12240
13240
14240
15960
16960
17960
18960

I'd like to configure the disks in the following fashion:

DiskSize (GB)Desired RAID Group
1960RAID10- DDB
2960RAID10- DDB
3960RAID10- DDB
4960RAID10- DDB
5960RAID10- DDB
6960RAID10- DDB
7960RAID5- Index
8960RAID5- Index
9300RAID1- Boot
10300RAID1- Boot
11240NA
12240NA
13240NA
14240NA
15960RAID5- Index
16960RAID5- Index
17960SPARE
18960SPARE

When I apply my Service Profile (SP) to the physical server, I get this error:

Failure Reason: There are not enough resources overall
Insufficient disks for the specified raid level
LUN Resource Selection Logs for Service Profile ecmsp09, LUN RAID5-INDX:

Order Description

1Disk selection process started for local lun: org-root/org-Dedicated/org-MedAgt/profile-STR-MediaAgnt01/das-scsi-lun-RAID5-INDX
2Try to find out an existing disk group for the new LUN
3Cannot carve out of the existing disk groups. Trying to create a new disk group
4Controller sys/rack-unit-8/board/storage-PCH-7 does not support OOB
5Failed to find sufficient disks

I don't get it. The disk groups should only use up 12 disks while there are 18 disks in the server.

I then removed the RAID5 disk group from the storage profile, and tried again, and get this error:

Failure Reason: There are not enough resources overall
Insufficient disks for the specified raid level
LUN Resource Selection Logs for Service Profile ecmsp09, LUN RAID10-DDB:

Order Description

1Disk selection process started for local lun: org-root/org-Dedicated/org-MedAgt/profile-STR-MediaAgnt01/das-scsi-lun-RAID10-DDB
2Try to find out an existing disk group for the new LUN
3Cannot carve out of the existing disk groups. Trying to create a new disk group
4Controller sys/rack-unit-8/board/storage-PCH-7 does not support OOB
5Failed to find sufficient disks

My disk groups look like this:

<lstorageDiskGroupConfigPolicy
childAction="deleteNonPresent"
descr="Storage for OS using 2 disks in RAID1"
dn="org-root/disk-group-config-STR-OS-2D-R1"
intId="80455880"
name="STR-OS-2D-R1"
policyLevel="0"
policyOwner="local"
raidLevel="mirror">

<lstorageVirtualDriveDef
accessPolicy="platform-default"
childAction="deleteNonPresent"
driveCache="platform-default"
ioPolicy="platform-default"
readPolicy="platform-default"
rn="virtual-drive-def"
security="no" stripSize="platform-default" writeCachePolicy="platform-default"/>

<lstorageDiskGroupQualifier
childAction="deleteNonPresent"
driveType="SSD"
minDriveSize="290"
numDedHotSpares="unspecified"
numDrives="2"
numGlobHotSpares="unspecified"
rn="disk-group-qual"
useJbodDisks="no"
useRemainingDisks="no"/>

</lstorageDiskGroupConfigPolicy>

<lstorageDiskGroupConfigPolicy
childAction="deleteNonPresent"
descr="Storage Group for RAID10 from 6 disks"
dn="org-root/disk-group-config-STR-DATA-6D-R10"
intId="80711146"
name="STR-DATA-6D-R10"
policyLevel="0"
policyOwner="local"
raidLevel="mirror-stripe">

<lstorageVirtualDriveDef

accessPolicy="platform-default"
childAction="deleteNonPresent"
driveCache="platform-default"
ioPolicy="platform-default"
readPolicy="platform-default"
rn="virtual-drive-def"
security="no"
stripSize="platform-default"
writeCachePolicy="platform-default"/>

<lstorageDiskGroupQualifier
childAction="deleteNonPresent"
driveType="SSD"
minDriveSize="950"
numDedHotSpares="unspecified"
numDrives="6"
numGlobHotSpares="unspecified"
rn="disk-group-qual"
useJbodDisks="no"
useRemainingDisks="no"/>
</lstorageDiskGroupConfigPolicy>

I won't post the RAID5 config unless it's required, but it's a 4x SSD disk RAID5 setup.

I then re-applied the Service Profile using just the "BOOT-R1" & "RAID10-DDB" disk groups and it looks like it would be successful, however it's using the wrong disks for the BOOT LUN:

Service Profile ecmsp09 (org-root/org-Dedicated/org-MedAgt/ls-ecmsp09) [Server: sys/rack-unit-8]

LUN Resource Selection Logs for Service Profile ecmsp09, LUN BOOT-R1:

Order Description

1Disk selection process started for local lun: org-root/org-Dedicated/org-MedAgt/profile-STR-MediaAgnt01/das-scsi-lun-BOOT-R1
2Try to find out an existing disk group for the new LUN
3Cannot carve out of the existing disk groups. Trying to create a new disk group
4Controller sys/rack-unit-8/board/storage-PCH-7 does not support OOB
5Select normal disk in slot: 7
6Select normal disk in slot: 8


LUN Resource Selection Logs for Service Profile ecmsp09, LUN RAID10-DDB:

Order Description

1Disk selection process started for local lun: org-root/org-Dedicated/org-MedAgt/profile-STR-MediaAgnt01/das-scsi-lun-RAID10-DDB
2Try to find out an existing disk group for the new LUN
3Cannot carve out of the existing disk groups. Trying to create a new disk group
4Controller sys/rack-unit-8/board/storage-PCH-7 does not support OOB
5Select normal disk in slot: 1
6Select normal disk in slot: 2
7Select normal disk in slot: 3
8Select normal disk in slot: 4
9Select normal disk in slot: 5
10Select normal disk in slot: 6

Is there a way to achieve the required result?

My last hope will be to go to the datacenter and re-arrange the disk order then modify the disk groups by setting them to "Disck Group Configuration (Manual)" and manually selecting the disk slots to use for each RAID group.

4 Replies 4

marce1000
VIP
VIP

 

 - These bug reports might be involved : https://bst.cloudapps.cisco.com/bugsearch/bug/CSCuy86027 and or https://bst.cloudapps.cisco.com/bugsearch/bug/CSCvd26834

 M.



-- ' 'Good body every evening' ' this sentence was once spotted on a logo at the entrance of a Weight Watchers Club !

Doesn't look relevant as one talks about creating LUNs using external means and the other is about using PCH controller definitions. But thanks for the info.

Steven Tardy
Cisco Employee
Cisco Employee

These UCS errors are very annoying:

There are not enough resources overall

When disks have this error most of the time the disk is NOT in the correct state to be consumed (but doesn't tell you that in the error message).

UCS RAID configuration will only use "Unconfigured Good" disks (but doesn't say that explicitly.)

Usually right-click the individual disks and set to "Unconfigured Good".

Hope that helps.

I've found that when working with servers with existing disk configurations, I need to delete the Orphaned LUNs else new Service Profiles will not apply. I have done this with the above server.

I've ensured that all physical disks on the server are in the Unconfigured Good state.

The Disk Groups & Service Profiles apply OK to the servers, as long as I remove the Minimum Drive Size value, however the wrong disks are used for each RAID group.

For now, I've asked the DC team to re-arrange the disks so that the 300GB disks are in slots 1 & 2 and the 960GB in the rest and I'll just create a manual slot assignment in the Disk Group. I don't think there is a way for UCSM to choose the correct disks in this situation!?

 

Review Cisco Networking products for a $25 gift card