Trouble creating a RAID array with VMWare ESX3i
HP DL140 Server, LSI Logic SAS1068 RAID controller, RAID, RAID 1, RAID Array, Raid Controller, VMware ESX, VMware ESX 3i
Hi,
We’ve set up VMWare ESX3i on a single disk on our HP DL140 Server. We’re now trying to set up a RAID array, using the in-built LSI Logic SAS1068 RAID controller. Unfortunately, it says that there’s not enough space at the end of the disk, and it would need to be formatted, in order to create the array. There is over 100 Gb available, so deleting something from the disk is unlikely to make any difference.
Does anyone have any experience with this or anything similar? Does it have anything to do with the way VMWare partitions the disk, (ie not NTFS or FAT)?
Thanks for your help,
Keith
Software/Hardware used:
Software/Hardware used:
ASKED:
January 8, 2009 2:19 PM
UPDATED:
January 16, 2009 5:01 PM
Answer Wiki:
In order to create a new RAID array the RAID controller has to have disks with no data on the. To add this disk to a new array you will need to remove the current file system from it deleting all the data. To keep your vmdk files (and other files for the virtual machine) you will need to move the files to another volume or server then delete the file system from the disk, create the array, reinstall VMware and move the vmdf files (and other files for the virtual machine) back. You can then add the virtual machines to the new VMware OS.
To see all answers submitted to the Answer Wiki: View Answer History.
The reason it says there is not enough space at the end of the disk is because it wants to create raid system area at the end of the disk, but this is currently taken up by your current file system, so as Mrdenny says you will have to backup all the data on the disk, and then delete the filesystem, the RAID controller will then be able to create the area it needs and then when you restore your data, you will find the filesystem is just a teeny bit smaller
Well, ok, it was clear that it wasn’t going to be possible to RAID the disc after it had been created. We have an identical server that we use as a spare / testbed. So I decided to move / convert all of the machines across gradually. The plan was then to move the discs back to the original server once done. This didn’t go quite according to plan.
Last night I moved the discs, then realised that both servers had lost all network visibility on all virtual machines, as well as the management network on the primary. (Incidentally, I noticed that any machines connected via the secondary NIC on a different subnet were visible, and completely fine.) After spending about 4 hours at it, I ended up resetting all settings to defaults. This really wasn’t good. It meant having to rebuild all of the network connections again from scratch. The virtual machines were in there quite happily, just completely invisible.
Today, after lots of research, I finally cracked the problem. Here’s how I solved it:
· At the console, Press Ctrl+F1. This takes you to the command line.
· Type unsupported
· You will receive a warning. Type the admin password.
· You will now be logged in locally.
· Type esxcfg-vswitch –l. This will show the status of the switching. You should see 2 different virtual switches, aka vSwitch0 & vSwitch1. (Note: the linux is case sensitive. If you type vswitch0, it won’t be recognised.) You will also see the relevant Port Group Names. There is an Uplinks heading against each virtual switch, and against each Port Group Name. If any of the Uplinks are blank, they will not be visible. (They should have vmnicX assigned to them.) In particular, if the Management Network, and the Internal LAN are blank, you won’t be able to manage the server remotely, or reach any of the virtual machines on the management network.
· To fix this problem, type esxcfg-vswitch –U=vmnic0 vSwitch0 (or esxcfg –unlink=vmnic0 vswitch0). This disconnects the vmnic0 interface from vSwitch0.
· Then type esxcfg-vswitch –L=vmnic0 vSwitch0 (or esxcfg –link=vmnic0 vswitch0). This re-connects the vmnic0 interface to vSwitch0.
· Type esxcfg-vswitch –l. Check that vmnic0 has indeed been assigned correctly this time.
· If there’s still a problem, try using esxcfg-vswitch –add-pg-uplink=vmnic0 –pg=(Port Group Name) vSwitch0.
· Once finished, type exit, and then press Ctrl+F2 to go back to the GUI.
Not sure if this is a common problem or not, but just in case, I thought I’d share it.
Thanks for your help,
Keith