Dl380 G6 - Sas Expander Issues

[d3vnul]

Hello,

Today we just installed a SAS Expander on a DL380 G6 , the server has the following disk setups:

Controller: P410i (firmware 3.66) with 512MB and Battery + sas expander

Both Cages have the same firmware (1.14 )

ilo version: 2.09

Total number of Disks installed on the server: 13

Cage1:
bay 1-4 -> RAID-10 ( 4 x 300GB SAS) (1st array)

bay 5-8 -> RAID-10 ( 4 x 300GB SAS) ( 2nd array)

note: bay1 disk was replaced today with a spare as the disk failed (it was replaced right before shuting down the server to install the sas expander and second 8sff drive cage).

Cage2:
bay 1-4 -> Unconfigured ( reserved to create a new RAID-10 ) ( 4 x 300GB SAS)

bay 5 -> Unconfigured ( 1 x 300GB SAS) -> reserved/unused-> keeping it as a replacement if any other disks will fail

bay 6-8 -> Empty

Now here are the problems we started to face:
After installing the sas expander and the second drive cage i seen the following log from IML:
POST Error: 1724-Drive Array - Physical Drive Position Change(s) Detected ( no idea why this happened)

POST Error: 1787-Drive Array Operating in Interim Recovery Mode (this should be related to the only replaced disk from Cage1/Bay1 as it was failed).

Short story the server should be like this:
Disk 1-4 -> 1st RAID-10 Array

Disk 5-8 -> 2nd RAID-10 Array

Disk 9-12 -> 3rd RAID-10 Array (not created yet)

Disk 13 -> Standby disk/Not to be used yet.

However take a look on what ESXi is telling about disk setup (check attachment). Nothing makes sence for me on how everything changed in a very weird way.

From the attached picture , in normal conditions it should be like the following:
Logical volume 1 should be the 1st RAID-10 Array consisting of Disk 1,2,3,4

Logical volume 2 should be the 2nd RAID-10 Array consisting of Disk 5,6,7,8 ( but it say it's from 6,7,8,9 loool ? )
And for some reason ESXi detects as there would be 14 disk onto the server while there are only 13 installed !
The problem is that Disk 1 shows as MISSING inside ORCA and it's being detected as Disk5

I wonder if anyone would/could understand all what i just said and if anyone also have some ideas on what to do or what happened ?

Thank to everyone in advance !

[Renjiv]

Hi,

Please share the ADU report from the server so that we can check the location change of the disk.

Regards,

Renji V

[d3vnul]

Hello,

Will try to provide it later on today , server is on production and we cannot reboot it to boot on the HPACU on working hours (this is also because we are under esxi. Will do this after 01AM when we can perform this task.

Regards.

[Dave73]

 (it was replaced right before shuting down the server to install the sas expander and second 8sff drive cage).

Did you allow the re-build to finish before you shutdown and installed the expander ?

[d3vnul]

We haven't because we were planning to delete and create the array again since we already migrated all data from that affected array. So we though it's faster to just delete it and create again.

Even so , i haven't noticed anything that indicated a rebuild has started.

[d3vnul · Accepted solution]

Update: I have just deleted the first affected array (the one with a failed disk that received a replacement) and after i created the array again everything seems to be in perfect order now. Disks order now shows properly.

Hope it will help someone that might face the same problem.

With best regards,

Alexandru.