Failure in Mapping or Unmapping a Volume on a VBS Node

Hi team, here's a new case.

Problem Symptom

When a volume is mapped to or unmapped from a VBS node, a failure message is displayed on DeviceManager.

Problem Diagnosis

1. Search for the vbs_cli log printed when the problem occurred to check whether the time spent on disk adding or deletion in the SCSI kernel exceeds 27 seconds. If the time exceeds 27 seconds, go to 1. If no vbs_cli log is found, go to Problem Diagnosis.

   The following is an example of the log:

   [root@HN_0_1 run]#grep -E "cli_ioctl_rmv_disk|cli_ioctl_add_disk" /var/log/dsware/vbs/vbs_cli/log-VBS_cli.log.0[06 March 11:14:23][vbs_cli.c,8134] cli_ioctl_rmv_disk[vol_name=131075]: cli ioctl RMV DISK [65535,2790656,14,0] succeed, spending time 35 s.
   

2. Analyze the dsware_vbs log printed when the problem occurred to check whether the time spent on disk adding or deletion in the SCSI kernel exceeds 27 seconds. If the time exceeds 27 seconds, go to 1. If no dsware_vbs log is found, this problem is caused by other reasons. In this case, go to 2.

   The following is an example of the log:

   [root@HN_0_1 run]#zcat /var/log/dsware/vbs/bak/log-dsware_dsware_vbs_272.log.2020-03-06_11-03-51.818.tgz | grep -aE "dsw_scsi_remove_device|scsi_add_disk"[2020-03-06 03:00.046][15000029a0000][WARN][11603][scsi_command.c,154,dsw_scsi_remove_device][tid:0x7f1288c75700][vbs:131075]sync Remove disk [2790656:65535:14:0] DONE. spending time 31 s.
   

Causes

When a volume is mapped to or unmapped from a VBS node, the SCSI kernel adds or removes disks. A disk can be removed from the SCSI kernel until the asynchronous detection I/Os of all added disks are returned. In a slow I/O scenario, the asynchronous detection I/Os are processed slowly, which prolongs the time for disk adding or removal in the SCSI kernel. If the processing time exceeds 27 seconds, the VBS node fails to map or unmap volumes due to timeout.

Solution

1. Map or unmap a volume to or from the VBS node again and check whether the operation is successful.

   If yes, no further action is required.

   If no, go to 2.

2. Contact technical support engineers.