Fault symptom: The disk mapped in the host can't be read and write, and show error: "disk is full" & "the file system structure on the disk is corrupt and unusable".
Host version:WindowsServer2008 x64
Windows alarm at first time:
windows alarm turns to:
1. Check the status of issued LUN of this host:
The LUN mapped to issued file system is“LUNACME-Input-”，and this LUN belongs to“ RAIDSATA_7”，status is normal. And we also checked there is no Bad Sector Block in such RAID.
2. RAID and disk status confirm
Member Disk List:
All disks status in such RAID are Normal.
3. IO read/write status confirm：
According to storage logs, there is no such IO error record。
Conclusion：According to above analysis, there is no storage error will directly cause such LUN file system corrupt, so we do analyze in host side.
1. Check Windows event information, we found in 2/15/2016 13:52(UCT +04), the host file system already corrupt, as below shows:
This alarm shows the volume ACME input already corrupt and unusable, system show the alarm "ACME to run chkdsk utility to repair file system". As volume F still could be accessed, so customer didn’t repair such volume.
2. Until 19th of February, whole volume F file system been damaged, now system notice: The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume F:, as below shows:
As the file system of whole volume F had been damaged, so we can’t do any read/write operation to this disk, and even get capacity information, as below shows:
3. And according to storage log analysis, there is no abnormal from storage could cause LUN data damaged, so the cause reason should be in host or link.
In storage log, we found the iSCSI link unstable, the link from host to storage frequently break down, as below shows：
4. Check the map relationship of issued LUN, the issued LUN（ACME-Input-）mapped to host 0, as below shows:
This wwn of host 0 is: d
Check the host break down record as below：
Controller A break down record：
Controller B break down record：
There is no abnormal alarm in storage logs when break down happen, should confirm the probable reason from host side.
5. Except the issued LUN host, other hosts also happen such break down records:
6. Check host system log event, we could find many initiator failed to target issues, as below shows:
The root cause of this break down should be confirmed by host side.
Note: file system repair may cause some files damaged.
1. Need stop operate issued disk when system report an error.
2. Need find the OS provider to check the OS system structure issue together