Problem
After an event of some kind (it was an array failure in our case, but the internet is awash in tales of power failures) all the VMs in my ESXi 4 system come up as "Unknown (Inaccessible)".
Examining the logs shows errors like:
Apr 17 22:10:18 gw-esxi-01 Hostd: [2012-04-17 22:10:18.742 FFC61E80 info 'App'] File path provided /vmfs/volumes/4da47646-42f22ed8-ffb8-001d0968d057/hsphere.w.t1/hsphere.w.t1.vmx does not exist or underlying datastore is inaccessible: /vmfs/volumes/4da47646-42f
Apr 17 18:10:20 gw-esxi-01 22ed8-ffb8-001d0968d057/hsphere.w.t1/hsphere.w.t1.vmx
Solution
- enable the SSH access on the console by going to Troubleshooting -> then enable the Tech SSH.
- Run esxcfg-volume -l and if it comes back as mountable, then get the datastore UUID or label. Then run esxcfg-volume -M $LABEL to mount it.
Example: datastore1 is not mounted.
~ # esxcfg-volume -l
VMFS3 UUID/label: 4da47646-42f22ed8-ffb8-001d0968d057/datastore1
Can mount: Yes
Can resignature: Yes
Extent name: naa.6001e4f017e810001720a4f308f790fb:3 range: 0 - 471295 (MB)~ # esxcfg-volume -M datastore1
Persistently mounting volume datastore1~ # df
Filesystem 1k-blocks Used Available Use% Mounted on
visorfs 1341308 267120 1074188 20% /
vfat 4192960 2624 4190336 0% /vmfs/volumes/4da47646-180cf51c-96bd-001d0968d057
nfs 206693144 95781064 100243316 49% /vmfs/volumes/b386bf1f-9f72248c
vfat 255716 4 255712 0% /vmfs/volumes/e1ab2e71-6e9d4ae8-3087-162dcd3f0867
vfat 255716 91340 164376 36% /vmfs/volumes/e97c7f73-1e33bc5b-112c-515d5c8f0643
vfat 292752 144096 148656 49% /vmfs/volumes/2da668ef-40e5d96b-90bf-855ddb9c5547
vmfs3 482607104 451988480 30618624 94% /vmfs/volumes/4da47646-42f22ed8-ffb8-001d0968d057
If it isn't mountable, google around for how to check/repair the partition table before mounting.