Back-End SAS Isolation

Hello, dear!

Have a nice day!

You can use the following method to locate the error:

1. Some back-end SAS links are isolated or an alarm is generated indicating that some back-end SAS links are isolated. (Note: Some isolation alarms do not cause back-end link down.)

2. Collect storage system logs and check the SES log SES_log_xxxxxxxxx_mem.txt of the controller at the point in time when the link is intermittently disconnected. Search for the keyword err inc, as shown in the following figure.

If a large number of err inc messages are printed in the log, bit errors occur on the port of the subrack.

Search for the keyword ERR DISABLED.

According to the SES logs, the SAS port has four isolated PHYs. If some PHYs are isolated, an isolation alarm is reported. If all PHYs are isolated, a link down alarm is reported and an isolation alarm is reported.

Run the showportswitch command in developer mode to query the PHY status of the SAS port, as shown in the following figure.

Hope can help you!

Dear, Jiro!

If your storage has many bit error sources on the link. Maybe you need to replace the component to local the fault.

1. Replacing a SAS Expansion Cable

2. Run the showportswitch command to check the PHY of the isolated port, and run the chgportswitch command to manually enable the isolated port. For details about the command format, see the appendix. (Note: After the SAS port is isolated, the storage control chip cannot detect the port change event. Therefore, the port must be manually enabled.)

3. After a period of time, collect storage logs and check whether bit errors (err inc) still exist on the link. If no bit error exists, the problem is solved. If bit errors still exist, go to the next step.

4.  Replace the SAS expansion module, interface module, and controller in sequence. 

Hope can help you!