resizing ESX 3.5 Linux partitions

[Ben Kevan]

On Monday 22 September 2008 11:24:32 pm sunhux G wrote:
> Hi,
>
> On my development VMWare ESX 3.5, I have the following :
> # fdisk -l
> Disk /dev/sda: 85.9 GB, 85905637376 bytes
> 255 heads, 63 sectors/track, 10444 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>    Device Boot    Start       End    Blocks   Id  System
> /dev/sda1             1     10444  83891366   fb  Unknown
> Disk /dev/cciss/c0d0: 440.3 GB, 440345714688 bytes
> 255 heads, 63 sectors/track, 53535 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>            Device Boot    Start       End    Blocks   Id  System
> /dev/cciss/c0d0p1   *         1       100    803218+  83  Linux
> /dev/cciss/c0d0p2           101      1120   8193150   82  Linux swap
> /dev/cciss/c0d0p3          1121      1603   3879697+  83  Linux
> /dev/cciss/c0d0p4          1604     53535 417143790    f  Win95 Ext'd (LBA)
> /dev/cciss/c0d0p5          1604      1985   3068383+  83  Linux
> /dev/cciss/c0d0p6          1986      2307   2586433+  fb  Unknown
> /dev/cciss/c0d0p7          2308      2320    104391   fc  Unknown
> How can I take some space from c0d0p4 to add it to c0d0p5?
> Is c0d0p5 & c0d0p6 (& what about c0d0p7) linked to each other?
> (As there's a +  sign there - what does the +  sign indicate).
> How can I mount c0d0p5 - it hangs  when I tried to mount it by
> "mount /dev/cciss/c0d0p5  /mount_directory".  Now when I issue
> "fdisk -l", the display lists till c0d0p7 & hangs there (doesn't come
> back to Unix command prompt even after the ESX is rebooted).
>
> I'm trying to a bit more space to the Linux fs below:
> # df -k
> Filesystem           1K-blocks      Used Available Use% Mounted on
> /dev/cciss/c0d0p3      3818672   1223876   2400812  34% /
> /dev/cciss/c0d0p1       790556     38440    711956   6% /boot
> none                    134284         0    134284   0% /dev/shm
> /dev/cciss/c0d0p5      3020140    267092   2599632  10% /var
>
>
> I'm comparing with our Production ESX 3.5 :
> Disk /dev/sda: 75.1 GB, 75169267712 bytes
> 255 heads, 63 sectors/track, 9138 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>    Device Boot    Start       End    Blocks   Id  System
> /dev/sda1             1      9138  73400921   fb  Unknown
> Disk /dev/cciss/c0d0: 146.7 GB, 146778685440 bytes
> 255 heads, 63 sectors/track, 17844 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
>            Device Boot    Start       End    Blocks   Id  System
> /dev/cciss/c0d0p1   *         1        13    104391   83  Linux
> /dev/cciss/c0d0p2            14       650   5116702+  83  Linux
> /dev/cciss/c0d0p3           651     17508 135411885   fb  Unknown
> /dev/cciss/c0d0p4         17509     17844   2698920    f  Win95 Ext'd (LBA)
> /dev/cciss/c0d0p5         17509     17577    554211   82  Linux swap
> /dev/cciss/c0d0p6         17578     17831   2040223+  83  Linux <== only
> one with +
> /dev/cciss/c0d0p7         17832     17844    104391   fc  Unknown
> & we have bigger Linux fs :
> Filesystem           1K-blocks      Used Available Use% Mounted on
> /dev/cciss/c0d0p2      5036316   1296988   3483496  28% /
> /dev/cciss/c0d0p1       101089     26276     69594  28% /boot
> none                    134284         0    134284   0% /dev/shm
> /dev/cciss/c0d0p6      2008108     63180   1842920   4% /var/log
>
>
>
>
>
> I have the following steps from someone to resize a Linux swap partition.
> It gave no
> error when the steps/commands are executed (in Unix Single user mode) in
> ESX 3.5
> but after it's rebooted, the new size did not take effect, thus it did not
> work.  I probably
> can't use it this time to resize this time :
>
> use "fdisk -l" to find out if swap device ie /dev/hda3
> parted
> select /dev/hda3
> print
> resize  START END
> quit
> quit

I'm trying to figure out what you're really trying to do, and why? 

why do you need to increase the disk space to the "Linux File System" and what 
Mountpoint are you trying to increase? 

I have about 25 ESX boxes, with the default disk layout, and have never had a 
problem (although on some dev boxes, I've changed to allow more for / so I 
can develop scripts for patches). 

Are you trying to increase swap? If so, Why? if you're filling up swap you 
have other issues, and it isn't the swap size. ESX 3.5 is much better with 
swap since most of the virtualization overhead is now sent to vmkernel and 
not to the sc. 

ben