No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

An HBA Issue Led to a Failure to Fail Back After an Optical Fiber Was Removed and Reinserted

Publication Date:  2013-08-08 Views:  41 Downloads:  0
Issue Description
Product and version information:
  • S5500T V100R001 V100R002
  • S5600T V100R001 V100R002
  • S5800T V100R001 V100R002
  • S6800T V100R001 V100R002
  • Application server operating system: SUSE Linux Enterprise Server 11 (x86_64)
  • HBA: QLogic HBA

The storage device was properly connected to the host through a direct connection redundant network. A LUN was mapped to the host and was owned by controller A. The host performed I/O operations on the LUN. During the operations, the optical fiber for controller A was removed, and the I/O path switched over to controller B. After the optical fiber was reinserted, the I/O path did not fail back to controller A.
Alarm Information
None
Handling Process
  1. Remove and reinsert the optical fiber again. Then run the hot_add command on the host to scan for LUNs, and the upTools -S command to check whether any new paths are generated.
  • If yes, the handling procedure is complete.
  • If no, go to 2.
      2.    Check whether the optical fiber is damaged. If no, remove the optical fiber and insert it to another port, or add another fiber connection. If yes, replace the optical fiber. Run the hot_add command on the host to scan for LUNs, and the upTools -S command to check whether any new paths are generated.
  • If yes, the handling procedure is complete.
  • If no, go to 3.
       3.     It is recommended that the services be migrated to another host and the current host be restarted. After the system has recovered, migrate the services back.
Root Cause
Background: When the host was performing I/O operations on the LUN, the optical fiber for controller A was removed, and the I/O path switched over to controller B.
After the optical fiber was reinserted, we ran the upTools -S command, only to find out that no path was generated on controller A. The output is shown in the following figure.

By viewing the information in /var/log/messages, we found out that no physical device was reported, and the system could not generate a physical path. If any device was reported, information similar to the following figure would have been displayed.

Therefore, the I/O path could not be failed back to controller A. Based on the preceding analysis, it is concluded that an HBA driver issue led to a failure to report physical devices.
Suggestions
Do not unnecessarily remove or reinsert optical fibers.

END