Summary: If VxRail appliance have external iSCSI storage attached, after performing graceful shutdown and power on all the nodes, service VMs are not able to start up automatically, and all the other nodes stay in maintenance mode except node 1.
Article Content
| Issue | If VxRail appliance have external iSCSI storage attached, after performing graceful shutdown and power on all the nodes, service VMs are not able to start up automatically, and all the other nodes stay in maintenance mode except node 1. Look into the syslog.log on each node, similar error log could be found as below: 2017-05-25T11:32:39Z startSystemVM: 2017-05-25 11:32:39,154.154Z - ERROR Failed to autostart VM's due to: Traceback (most recent call last): File "/etc/rc.local.d/998.start_vm.py", line 355, in <module> main() File "/etc/rc.local.d/998.start_vm.py", line 320, in main config = load_config_file(conn) File "/etc/rc.local.d/998.start_vm.py", line 45, in load_config_file with open(os.path.join(VOLUME_PREFIX, ds.name, CONFIG_FILE_PATH), 'r') as fp: IOError: [Errno 2] No such file or directory: '/vmfs/volumes/Unity-Data02_6T/images/start_vm.conf' <<=== Unity-Data02_6T is the external iSCSI storage |
| Cause | During the node start up phase, VxRail is trying to find out SATADOM by seeking a datastore with type VMFS. If external iSCSI storage attached, VxRail may mis-recognize external iSCSI storage as SATADOM since they have the same type VMFS. Since service VMs are stored in SATADOM, VxRail will be failed to power on the service VMs on the external iSCSI storage. |
| Resolution | Temporary detach the iSCSI storage from VxRail appliance (Power off the external iSCSI storage or disconnect it from VxRail), then power on all VxRail nodes. Re-attach iSCSI storage to VxRail when it is fully started up. |