Oracle Exadata Inspection Method

Automatic Inspection Method One

1. Collect Exachk reports

(Execute the following commands on the first compute node)

# unzip /tmp/AHF-LINUX_v26.3.1.zip

# /tmp/ahf_setup

# exachk

Note:

Download the latest version of AHF from MOS: (KB858617) Autonomous Health Framework (AHF) - Including Trace File Analyzer and Orachk/Exachk

Starting from X8M-2, executing the exachk command will prompt for the ROCE switch address. For example: dm02sw-rocea0,dm02sw-roceb0

Manual Inspection Method One

1. Collect ILOM logs

(Execute the following commands on the first compute node)

[root@exa-db01 ~]# vi /tmp/all_group

Write the hostnames of all compute nodes and storage nodes into this configuration file, with one entry per line.

[root@exa-db01 ~]# vi /tmp/dbs_group

Write the hostnames of all compute nodes into this configuration file, with one entry per line.

[root@exa-db01 ~]# vi /tmp/cell_group

Write the hostnames of all storage nodes into this configuration file, with one entry per line.

2. Collect ILOM logs for compute nodes

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "ipmitool sunoem cli 'show faulty'"

--Collect ILOM logs for storage nodes

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "ipmitool sunoem cli 'show faulty'"

3. Collect storage software logs

(Execute the following commands on the first compute node)

--Collect storage software logs for compute nodes

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "dbmcli -e list alerthistory"

--Collect storage software logs for storage nodes

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list alerthistory"

4. Collect storage information

----Disks on storage nodes

(Execute the following commands on the first compute node)

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "df -Ph"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list physicaldisk"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list lun"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list celldisk"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list griddisk"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list flashcache detail"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list cell attributes flashCacheMode"

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list flashlog detail"

##View the number of disk access errors

[root@exa-db01 ~]# dcli -g /tmp/cell_group -l root "cellcli -e list physicaldisk detail" |egrep "name|deviceName|Count"

----Disks on compute nodes

(Execute the following commands on the first compute node)

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "df -Ph"

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "dbmcli -e list physicaldisk"

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "dbmcli -e list lun"

[root@exa-db01 ~]# dcli -g /tmp/dbs_group -l root "dbmcli -e list physicaldisk detail" |egrep "name|errOtherCount"

5. Check clock synchronization status

[root@exa-db01 ~]# dcli -g /tmp/all_group -l root date

6. Collect IB switch information (IB switches exist only before X8M; execute for each IB switch)

# ssh root@exa-iba01

[root@exa-iba01 ~]# showunhealthy

[root@exa-iba01 ~]# checkpower

[root@exa-iba01 ~]# showdisk

[root@exa-iba01 ~]# env_test

[root@exa-ibb01 ~]# version

7. Collect ROCE converged Ethernet switch information (ROCE switches exist starting from X8M; execute for each ROCE switch)

# ssh admin@exa-rocea0

exa-rocea0# show inventory

exa-rocea0# show environment

8. Collect Cisco switch information

# ssh admin@exa-adm01

exa-adm01> show inventory

exa-adm01> show environment