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

Device Names are Inconsistent at Different Nnode in the two-node AIX OS of the S6800E Multipathing Sstorage System

Publication Date:  2012-07-20 Views:  75 Downloads:  0
Issue Description
A two-node AIX OS is formed by using the SF. The LUN mapping sequence is changed on the storage device before the two-node AIX OS is configured. After the two nodes are restarted, a cluster system is deployed. The upadm show lun command is executed on the two nodes and find that the mappings between LUNs and the disks are inconsistent. The mapping cannot be changed by running the cfgmgr command.
Alarm Information
None
Handling Process
Run the rmdev -dl hdiskX/updisk command to delete the definitions of all block devices (hdiskX or updiskX) on the host with incorrect mappings from the Cudv library of the AIX OS, and then run the cfgmgr –l fscsiX command to scan disks and regenerate hdiskX or updiskX. In addition, when services are stopped, running the rmdev command to delete hdiskX or updiskX has no risk. If upper-layer applications are based on file system, inconsistent LUN mapping sequences have no impact on applications.
Root Cause
As the LUN mapping sequence is changed on the disk array, the LUN mapping sequence on each host may be changed. The specific mapping policy is implemented by the PCM KE driver of the AIX OS. What is the risk caused by the inconsistent mappings? In the two-node mode, if the LUN mapping sequences are inconsistent on two host, check whether upper-layer applications are based on block device (hdiskX or updiskX) or file system. If upper-layer applications are based on file system, inconsistent LUN mapping sequences have no impact on the applications. If upper-layer applications are based on block device, the actually used LUNs are not the mapped LUNs as the mappings between block devices and LUNs have changed. Therefore, applications encounter errors.
Suggestions
When an IBM microcomputer is connected to a Huawei Symantec storage array and a cluster needs to be configured, it is recommended to install the multipathing software and then run the cfgmgr command to scan disks. If the storage system has scanned disks and the LUN mapping sequence has been changed on the disk array, it is recommended to run the rmdev -dl fcsX –R command (fcsX indicates the HBA, X is obtained by running the lsdev -Cc adapter command) to delete the HBA from all cluster hosts, and then run the cfgmgr command on cluster hosts to scan disks, ensuring that all hosts in the cluster scan LUNs in the same sequence.

END