How to increase datastore on ESXi 3.5

When you added the additional drive, have you Grown the Array in the Dell PERC 5/i BIOS Utility?

If you hit the Increase Button, do you not get the option to increase the space on the datastore, i.e. select the remaining space?

Yes, okay, I can see the 544.50 capactity in the Datastore!

whoops! my eyes are dim!

As you can see in the screen shot the datacapacity is present within Dell PERC, just not sure how to add that extra space to Datastore.

Yes, my mistake, didn't see it.

When you Hit the Increase Button underneath Rename, do you get a dialogue box, which shows the remaing space on thew (LUN) disk, that you can select to grow the current datastore named datastore1.

It should show 141GB of extra (space) unallocated space?

No I do not get that.  Here is a screen shot of what is present.
Vmware1.gif

Additionally, if I go to Configuration, Storage, Add Storage, and select Disk/LUN no options there either.  I was hoping to at least add a new Datastore and possibly span.

It should show the remaing space on the disk (LUN), if available.

How was the Array expanded?

Have you checked with fdisk at the console, what partitions exist on the disk?

Because in the breakdown of partitions on that disk, there are a few DOS, Extended Partitions, Logical Partitons.

Was the disk new, before inserting, am just wondering if it cannot extend because there is no free space, only way to check using fdisk, or Gparted (boot cdrom).

Just view, do not edit.

I will double check the status of array tomorrow morning when I can down the production servers.  I will verify that all the disk space has been extended via Dell's PERC management.  Thanks.

check there are no strange partitions using the space. there should really only be one vmfs partition, datastore1.

Checked the space via the Dell RAID Management and the space is properly allocated.  I am showing an Array with 544.50GB of space.  I booted the server and attempted again within ESXi and get the same results as before.  No additional space to add...nor can I create a new volume.  Will continue web research hoping to locate an answer.

But when you view it with Non ESXi, e.g. boot from Gparted Live CDROM

http://gparted.sourceforge.net/livecd.php

what does the physical disk partition look like?

is it a single VMFS 400GB partition, or are they other partitions?

or check with fdisk in ESX. (at the service console)

also where did the "new disk" come from was it brand new, or from another server?

I will have to boot tomorrow morning with GPARTED and get back to you.  The disks where NEW when added.  Since this is ESXi 4.0 I don't have the options to check via command line using FDISK.  Thanks for the tip with GPARTED.  I'll post results tomorrow.

ESXi 4.0 does have a console, and fdisk should also work.

I cannot figure out how to get to a command line in ESXi 4.0.

From the console at the servers keyboard

1. From the ESXi console summary screen hit ALT-F1.
2. Enter the word “unsupported” (without quotes).
3. Enter in the root password for your system.

Awesome..new trick...tx.  OK....I ran FDISK and the partition shows the size to be 584.6GB.  

is the partition 584.6 or the disk?

did the Dell RAID Array re-size the partition from 400 to 584?

so if you complete a fdisk -l

The Dell RAID Manager shows the disk as the full capacity...ARRAY0 544.50GB.  At the ESXi 4.0 console, fdisk -l shows the local DELL Disk (naa.6001e4f02c6714000ffe6d9774f46f00) at 584.6GB.  So I am now completely perplexed as to why or how these disk sizes can be so different.  However, if you do look back at my screenshot of the disk partition in vSphere Client, the same the local DELL Disk (naa.6001e4f02c6714000ffe6d9774f46f00) is 544.50GB...same as the Dell RAID Manager.  Not sure why the fdisk -l shows a larger size.  Either way, unclear as to why I am unable to expand on the space.  

well this could be a rounding issue, of the use or 1000 of 1024k.

but the question is not the size of the disk. That is clearly now 584GB.

but what is the sizie of the vmfs partition on the disk, and where is the remaining space, it is allocated to something else?

what I'm referring to, is why is there Logical and Extended partitions being detrected on this disk, after the VMFS partition!

can you paste the output of fdisk -l

the output should look somethinf like this

http://www.cyberciti.biz/faq/linux-viewing-drive-partitions-with-fdisk-parted/

and if you pipe with fdisk -l > disk.txt

you can copy the disk.txt file, or type here.

Sorry for the ignorance but I am not a Linux guru at all.  I run the fdisk -l >disk.txt.  I then ran the ls command and I see the disk.txt output file.  Can I attach a USB jump drive and copy the disk.txt file to it?  How?

no, you cannot attach a USB drive to ESX and copy it.

you would need to remote into the server, from a workstation, and use WinSCP (free download), SCP is a secure version (SSH) version of FTP to download the file, or view the file from a PC.

have you ever SCP-ed to the server?

http://winscp.net/eng/download.php

WinSCP download here.

But if you've never SCP-ed or SSH-ed to the host before, we've got some more tricks to show you.

No problems with the Linux Guru, here to guide and help you out. That's what we do here at EE.

Never.  Looks like I'm fixing to learn.

What we need to is to establish, what partitions are currently on the disk. To hopefully save you time, in backing up and moving all the VMs off, deleting the VMs, and datastore, reformatting, moving all the VMs (restoring) back.

without this relevant disk information, and why we got odd partitions being displayed, at present it is difficult to advise, why the space is missing.

Ok.  WinSCP installed.  Have WinSCP Login Screen awaiting.  I'm assuming the hostname I should use the ESXi server IP, port # is defaulted to 22? Username is root?  password should be my password?  Private Key??? Protocol SCP?

That's correct, have you connected via SSH or SCP before to ESXi?

Beause by default root login is not allowed so we need to change some config on the ESXi server.

NO....Just found the following steps:

By default this isn’t possible. But there’s a way to get this working, just do the following:
 1. Go to the ESXi console and press alt+F1
 2. Type: unsupported
 3. Enter the root password(No prompt, typing is blindly)
 4. At the prompt type “vi /etc/inetd.conf”
 5. Look for the line that starts with “#ssh” (you can search with pressing “/”)
 6. Remove the “#” (press the “x” if the cursor is on the character)
 7. Save “/etc/inetd.conf” by typing “:wq!”
 8. Restart the management service “/sbin/services.sh restart”

Is this accurate?

yes, that's correct, we are modifying SSH to allow root login.

oh hang, on, too quick.....

better article here, because you've got to kill a process as well maybe

http://www.techhead.co.uk/vmware-esxi-how-to-enable-ssh-connectivity

you also may have to permit root to login as well after enabling SSH

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=8375637

from 3. onwards

otherwise when you try SCP, it will connect, but not allow root to login.

Worked like a champ.  I figured out the rest.  Here is the results of the disk.txt:


Disk /dev/disks/mpx.vmhba33:C0:T0:L0: 1031 MB, 1031798784 bytes
64 heads, 32 sectors/track, 984 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes

                           Device Boot      Start         End      Blocks  Id System
/dev/disks/mpx.vmhba33:C0:T0:L0p1   ?   1573910   1694359 123339962   78  Unknown
Partition 1 has different physical/logical beginnings (non-Linux?):
     phys=(518, 102, 15) logical=(1573909, 35, 30)
Partition 1 has different physical/logical endings:
     phys=(743, 0, 62) logical=(1694358, 47, 17)
Partition 1 does not end on cylinder boundary
/dev/disks/mpx.vmhba33:C0:T0:L0p2   ?    211363    590115 387841909+  10  Unknown
Partition 2 has different physical/logical beginnings (non-Linux?):
     phys=(205, 7, 0) logical=(211362, 54, 14)
Partition 2 has different physical/logical endings:
     phys=(920, 235, 50) logical=(590114, 45, 24)
Partition 2 does not end on cylinder boundary
/dev/disks/mpx.vmhba33:C0:T0:L0p3   ?    912873   1849997 959615034   8b  Unknown
Partition 3 has different physical/logical beginnings (non-Linux?):
     phys=(260, 125, 54) logical=(912872, 22, 4)
Partition 3 has different physical/logical endings:
     phys=(893, 46, 60) logical=(1849996, 25, 23)
Partition 3 does not end on cylinder boundary
/dev/disks/mpx.vmhba33:C0:T0:L0p4   ?   1989377   1993441   4161545+   a  OS/2 Boot Manager
Partition 4 has different physical/logical beginnings (non-Linux?):
     phys=(269, 111, 50) logical=(1989376, 0, 1)
Partition 4 has different physical/logical endings:
     phys=(0, 0, 0) logical=(1993440, 0, 19)
Partition 4 does not end on cylinder boundary

Partition table entries are not in disk order

Disk /dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00: 584.6 GB, 584652423168 bytes
64 heads, 32 sectors/track, 557568 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes

                                           Device Boot      Start         End      Blocks  Id System
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p1             5       900    917504    5  Extended
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p2           901      4845   4039680    6  FAT16
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p3          4846    418176 423250944   fb  VMFS
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p4   *         1         4      4080    4  FAT16 <32M
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p5             5       254    255984    6  FAT16
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p6           255       504    255984    6  FAT16
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p7           505       614    112624   fc  VMKcore
/dev/disks/naa.6001e4f02c6714000ffe6d9774f46f00p8           615       900    292848    6  FAT16

Partition table entries are not in disk order

okay, this is what is really odd...

can you see, p1, p2, p3 are the partitions at the start of the disk.

p3 is the VMFS large datastore of 484.

and then we have another p4, p5, p6, p7 and p8, and they shouldn't be there, and that's why you've no space.

I have no idea, why you have additional partitions, p5, p6, p7 & p8.

you could attemtp to remove these partitions, and ESXi 4.0, may then see the remaining space on the datastore.

BUT, I would highly advice before any attempts are made to remove partitions, that a backup of your datastore, and any VMs that are on your datastore are backed-up fully, before any attempts are made to make any MODIFICATIONS to the datastore partition table.

Modificiations to the partition could possibly destroy ALL data on the disk.

no sorry, ignore all that!!!!

it's rubbish, I'm reading the values wrong.

Let me explain what I have seen when compared to another ESXi installation.   This server was created with a single disk array in which ESXi was installed.  On our other system we created two arrays.  ESXi was installed in the first array and the second array became the datastore.  This is how I usually build an ESXi or ESX server as it keeps the OS seperate from the DATASTORE.

In this case all the partitions for ESXi live in the same array.  I suspected this could be the cause of the problem.  I compared both my servers and the same partitions are in both...even comparable in size except for the VMFS partition.  On the second server the DATASTORE I created on the second array only has a VMFS partition showing....the other partitions do not exist.

I guess I am at a point of needing to move my guests off this server and rebuild the server correctly with two arrays...one for ESXi and the other for the DATASTORE, then return the guests to the system?

Okay, I see the issue here, it's due to partition layout, and ESXi installation and VMFS datastore.
(which has 8 partitions already)

so you've already got 4 primary partitions, and that's the max.

So this is documented well here, with what you need to do.

http://kb.vmware.com/selfservice/documentLinkInt.do?micrositeID=&externalID=1007026

but my point about valid backups is important.

It's much easier if your OS and Datastores are separate, because you'll never get this issue.

But there's the KB document, you can work through to solve it quickly, without the move VMs off etc.

Up to you how you wish to proceed.

Thanks for all your assistance.  I will look at the TID and analyze my options.  Will post back tomorrow.  I appreciate all you knowledge transfer here in the forum.  

hanccocka,

After careful review and consideration...we have opted to wait on acquiring the additional space at this time. This is our production server and we do not want to risk loss...nor can we afford the downtime.  We will investigate options to move off all guests and then rebuild appropriately.   All your input was valuable and much appreciated.  Thanks for all your help.

Thanks.

This is an issue that can occur, if you install ESX on the same set of disks at a datastore (still relevant in ESXi/ESX 4.1), hence the recommendation to install on USB/SD card (if ESXi) or pair of RAID 1 disks, and then vmfs datastore on a separate set of RAID disks.