CUUC 8.6 RAID

[George Banuelos]

Hi,

There is a MCS7845I3-K9-CMD2 (CUUC 8.6) on site with 4 drives.  I understand per Cisco documentation that the MCS7845I3-K9-CMD2 (CUUC 8.6) has a dual RAID 1 consisting of 4 hard drives.  I have a TAC/RMA case opened due to drive 0 failing and have the replacement in hand.  The questions are these:

1) are these 4 drives copies of the same logical drive or is each set of RAID 1 drives different?

2) how many of the drives can fail before the CUUC crashes completely

3) I read a Cisco doc outlining the procedure for failed drive replacement mentioning the possibility of a failed rebuild when you introduce the new drive - can anyone please comment on any methods to avoid this...

4) the Show Hardware log shows that the raw size of the hard disks are 136.731 GB, but the drives are 146 GB - is this loss due to formatting?

Thank you!!

[Manish Gogna · Accepted solution]

Hi,

As per the following link

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/trouble/8_6_1/trbl861/tbsystem.html#wp1190377

Performing Failed RAID Disk Replacement Without Restart

Perform the following procedure to replace a failed RAID disk for these specific servers:

•MCS-7835-H2

•MCS-7835-I2

•MCS-7845-H2

•MCS-7845-I2

•MCS-7835-I3

•MCS-7845-I3

•DL-380-G6

Step 1 Log in to the console as an Administrator and enter the CLI command, show hardware.

Step 2 Check the status of the logical drives. Perform one of the following:

–If the logical drive status is OK or Optimal, you need not perform any further action.

–If the logical drive status is not OK, check the physical disk status, as described in Step 3.

Step 3 Enter the CLI command, show hardware, again and perform one of the following:

–If none of the physical disks displays the status as "Failed", you need not perform any further action.

–If the logical drive status is not OK or Optimal, and any physical disk displays as "Failed", identify the physical disk on the server as follows.

The LED color of the failed disk will be in amber or red.

Note You must perform step 2 to verify the logical RAID drive status and then perform step 3 to verify the physical disk status.

Step 4 Pull the failed disk from the slot.

Step 5 Enter the show hardware CLI command to ensure that the current number of physical disks is reported.

It reports:

–On a 7835 class of server, only one physical disk in the show hardware CLI command output.

–On a 7845 class of server, only three physical disks in the show hardware CLI command output.

Step 6 Replace the empty slot on the failed disk with a new disk that is of the same type, of same manufacturer, and of the same size as that of the original disk. For example, Western Digital.

Step 7 Ensure the new replacement disk is inserted all the way in.

Step 8 Run the show hardware CLI command to ensure that the newly inserted physical disk has been detected.

It reports:

–On a 7835 class of server, only two physical disks in the show hardware CLI command output.

–On a 7845 class of server, only four physical disks in the show hardware CLI command output.

Step 9 If the correct disks are not reported, pull out the new disk and repeat from Step 5.

Step 10 Enter the CLI command show hardware to check the RAID status. It displays the rebuild percentage completion.

Rebuild will take between 1 to 2 hours to complete. This depends on the size of the disk.

After the failed RAID disk replacement is complete, the status of both the logical drive and the new physical disk will display as "OK" and "Online".

HTH

Manish

[George Banuelos]

Hi Manish, thank you for your reply.  I've read this doc, but it doesn't provide insight into how many drives can fail before the system goes down.  It would be beneficial to understand what type of data is contained in each set of drives.  For example: does the 1st RAID 1 with 2 hard disks contain the OS and the 2nd RAID with 2 hard disks contain messages and config?  ... or ... is all data split across the dual RAID 1.

Do you know why these servers just don't come with large enough drives on a RAID 1 so that all 4 drives are identical - wouldn't his be more robust than what is in place now... any thoughts on this?

... also ... the Cisco documentation mentions that the MCS 7845I3 has a dual RAID 1, but it did not specify if the RAID is 1 + 0 or 0 + 1.  At first glance, it seems that the documentation presents it to be RAID 1 + 0, but a TAC engineer mentioned it was 0 + 1 - kindly clarify this if you know.

On a final note, the raw size shows up as 136.731 GB.  Just curious as to where the 10 GB loss happens - please comment if you know.  It may be helpful for documentation to mention that the 146 GB drives will show up as 136.731 on the Show Hardware - I did not get a clear answer from Tier 1 TAC on this.  I was assured that the drive that was provided as part of the RMA process would be compatible, but the difference in size representation through me off because Cisco's documentation clearly mentioned that only the exact size HD should be used.

Thank you!!

George