[lustre-discuss] lfs doesn't like non-sequential ost indices

[Chris Hunter]

Hi,
This is a minor point. I am adding OSTs to a lustre 2.5 filesystem. I'm 
working with 2 OSS servers: srv1 & srv2. srv1 hosts ost indices 0,2,4,6, 
srv2 hosts ost indices 1,3. We just added ost indices 4 & 6 and will add 
5 & 7 later.

I ran lfs command on a client. I found "lfs osts" gave correct output 
(ie. ost index 5 not listed)
> # lfs osts
> OBDS::
> 0: testfs-OST0000_UUID ACTIVE
> 1: testfs-OST0001_UUID ACTIVE
> 2: testfs-OST0002_UUID ACTIVE
> 3: testfs-OST0003_UUID ACTIVE
> 4: testfs-OST0004_UUID ACTIVE
> 6: testfs-OST0006_UUID ACTIVE

However "lfs df" command gave an error about ost index 5 missing eg)
> # lfs df -h
> UUID                       bytes        Used   Available Use% Mounted on
> testfs-MDT0000_UUID       371.5G       77.8M      346.7G   0% /lustre/testfs/client[MDT:0]
> testfs-OST0000_UUID        14.2T       23.6M       13.5T   0% /lustre/testfs/client[OST:0]
> testfs-OST0001_UUID        14.2T       24.7M       13.5T   0% /lustre/testfs/client[OST:1]
> testfs-OST0002_UUID        14.2T       23.6M       13.5T   0% /lustre/testfs/client[OST:2]
> testfs-OST0003_UUID        14.2T       23.6M       13.5T   0% /lustre/testfs/client[OST:3]
> testfs-OST0004_UUID        14.2T       11.6M       14.1T   0% /lustre/testfs/client[OST:4]
> OST0005             : Resource temporarily unavailable
> testfs-OST0006_UUID        14.2T       11.6M       14.1T   0% /lustre/testfs/client[OST:6]

This appears to be a minor reporting issue.

chris hunter
chris.hunter at yale.edu