Error: “I/O timeout fetching online state from disk”
- DxEnterprise 15.0
- DxEnterprise 15.5
- DxEnterprise 16.0
When failing over migrated disks that have been replicated between multiple sites, you may encounter an error stating “I/O timeout fetching online state from disk…”
Log Name: DH2i Source: DxStorMonitor Date: <date_time> Event ID: 7018 Task Category: None Level: Error Keywords: Classic User: N/A Computer: <hostname> Description: I/O timeout Fetching online state from disk <disk_number>, command: <?xml version="1.0" encoding="utf-8" ?><rss><action>ll-device-update-details</action><params><para><device_path></para></params></rss>
Log Name: DH2i Source: DxLMonitor Date: <date_time> Event ID: 3003 Task Category: None Level: Warning Keywords: Classic User: N/A Computer: <hostname> Description: Hostability alert for vhost <vhost> on <node>: Disk(s) unavailable: <disk_name>
When there is an I/O event (i.e. PnP event, get disk detail), DxStorMonitor issues an IOCTL_DISK_GET_DISK_ATTRIBUTES to check whether or not the disk is online. If the query takes longer than 15 seconds, the DxStorMonitor will re-read disk identity information and re-compute the disk ID. The computed disk ID should be part of the managed disk list through the disktag.
There is a defect in the disk association routine for any migrated or replicated disks, wherein the disktag is no longer associated with the newly computed disk ID. This defect was fixed in DxEnterprise version 17 and higher.
To work around this disktag re-association issue, please follow the procedure to reset the disktag for any migrated or replicated disk:
- Freeze the parent Vhost
- Un-manage all aliased disks associated with the Vhost
- Re-manage the disks un-managed in step (2)
- Associate the re-managed disks with the Vhost diskgroup
- Unfreeze the Vhost