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>Search

Reminder

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

upgrade
Knowledge Base

In ALUA mode, after recovering mirror trail, the trail in host can’t recover and IO disconnect later

Publication Date:  2012-10-18  |   Views:  198  |   Downloads:  0  |   Author:  SU1001048203  |   Document ID:  EKB1000017508

Contents

Issue Description

Product and version information
1. The storage system S5500T&S5600T&S5800T&S6800T&S2600T  V100R001、V100R002
2. The host system Mac OS 10.7.0
Use MAC OS 10.7.0 host, open the ALUA function in array, map a LUN to host in array, this LUN is belong to B controller, use ibs down simulate mirror trail fault in array. After a period of time, recover the mirror trail by ibs up, then examine the LUN link state in host, we find the route state of A controller is unavailable (it should be initiative and not optimize in common sense), B controller is initiative and optimize; We read and write in the mapping LUN, IO is normal, pull out the connection label between B controller and host, IO disconnect.

Alarm Information

na

Handling Process

After recovering mirror trail, test unit ready command will ram UA sign, the host doesn’t send command to examine the state of interface group, the state of interface group can’t update, it’s belong to the host policy problem.

Root Cause

1. When we use ATTO FC HBA card, apple host has test mechanism, it sends test unit ready command to feel the change of array by cycle (1s); use ibs down simulate mirror trail fault in array, stop LUN, test unit ready command will ram UA TADRV_SCSI_SNSCODE_ ASYMMETRIC_ACCESS_STATE_CHANGED sign, host sends Report Target Port Groups command to examine the state of interface group, the state of interface group will become UNAVAILABLE when LUN stop.
2. Use ibs down simulate mirror trail fault in array, LUN from start to stop, test unit ready command will ram UA TADRV_SCSI_SNSCODE_ ASYMMETRIC_ACCESS_STATE_CHANGED sign. Later the host doesn’t send Report target port groups command to examine the state of interface group to update trail state. The mapping LUN read and write, IO is normal, pull out the connection cable between B controller and host, because the interface group state of A controller hasn’t been updated, it’s still UNAVAILABLE, it leads IO disconnect.

Suggestions

We can ignore it because it belongs to the problem of host policy.