Isilon unable to detect new drive

Hi Damal,

What is the status of the drive?

By status I mean, [HEALTHY], [NEW], [REPLACE], etc.

You can check that out by :

     # isi_for_array -s isi devices | grep -v HEALTHY

Try using these commands for adding the drive:

     # isi devices -a add -d :

     # isi devices -a format -d :

If it still doesn't get added, try rebooting the node.

You can also refer to this: NL 400 Drive Replacement Guide

Cheers,

Anurag

The new drive installed, status shows up as empty. I am not sure if format is performed but i cant check on that.

We can drain the client connections and reboot the node, but reboot is not preferred. Actually the node rebooted by itself and not sure why it is

I believe that step is part of procedure that CE is following to replace the disk.

But i will double check with him.

About the node reboot, support determined its a bug which will be fixed in 7.1.x.x, reboot caused by node panics - page fault in kernel module

Hi Damal, what is the Issue ID for this (panic/reboot)?

Thanks

-- Peter

The bug number is -119807

Is this what you asked for Peter.

Also i checked with few of my other colleagues about Isilon not detecting the replacement disk until node reboot, and they never came across this. They are on 7.0.2.4 as well. Is that a know issue ?

Thanks Damal

Just checked, no sight of 119807 in the 7.1.0.1 MR release notes.

As for the replaced disk not being recognized,

this has happened twice for us, but on 6.5 and IQ108NL.

In one case, just another drive was shipped that worked.

In the other case, we tried out three drives without luck.

Replaced the controller, still no drives worked in the particular slot.

Finally the whole chassis got replaced.

Cheers

-- Peter

Is the model of the new drive the same as the old drive?  EMC has been known in the past to change drive manufacturers and then you find out your drive is no longer on the supported list for the firmware you're running.

Double-check the node and drive firmware packages you have installed and update if necessary (which unfortunately requires node reboots). 

When the node rebooted we actually had disk failures in two different nodes - one in each node and after drive replacements, both the nodes has to be rebooted to detect the new drives. So i am not sure if it a s/w or h/w fault but i noticed its happening in my environment only - OneFS 7.0.2.4.

The chassis replacement you mentioned about, were you on 7.0.2.4 ?

Not aware of drive model but i can check with the CE. Is there a way to check the drive model and manufacturer from CLI ? 

The current firmware is from image below. Is there place where i can check the drive compatibility with the firmware on node ?

You can check the drive models with:

isi_radish -q

This also lists the firmware of the individual drives.  Go to the EMC support site, grab the drive release notes PDF, and see what's supported.

You can compare the drive models with the physical label of the drive your cluster won't see.  Your node firmware package looks recent.

> The chassis replacement you mentioned about, were you on 7.0.2.4 ?

6.5.5.11

-- P.

Ive gone through the docs

My understanding is we should have firmware revision: SN03 for drives, but they have SN02 which may not be compatible.

Bay 16/da26   is ST4000NM0033-9ZM170 FW:SN02

Bay 17/da8    is ST4000NM0033-9ZM170 FW:SN02

Bay 18        is not a recognized drive

Bay 19/da9    is ST4000NM0033-9ZM170 FW:SN02

Bay 20/da27   is ST4000NM0033-9ZM170 FW:SN02

There is a known issue related to drive add resulting in an 'empty' status.  Bug# 122361.  I would recommend you open an SR to validate if this is what you are hitting and the expected fix date for this issue.