I already explained how to resize an LVM online in one of my previous post. (http://ermanarslan.blogspot.fr/2013/10/linux-extend-logical-volume-size-lvm.html)
In that post, I gave all the info for resizing LVM based mount points. The explanation was from stratch. (scan scsi bus, fdisk, resize2fs and so on) and I have followed that approach many many times.
However, today, when I was working on a EBS application tier environment which was running on a RHEL 4 32 bit server, I noticed something and I want to share it with you.
I was requested to enlarge a mount which was based on LVM and which was mounted to root (/).
Again, I followed the same approach as documented in the blogpost, pointed by the above url but this was when I was enlarging the filesystem, the following error produced
[root@erpapp ~]# cat /etc/redhat-release
Red Hat Enterprise Linux AS release 4 (Nahant Update 5)
[root@erpapp ~]# resize2fs /dev/mapper/Cluster1-Root
resize2fs 1.35 (28-Feb-2004)
/dev/mapper/Cluster1-Root is mounted; can't resize a mounted filesystem!
So, it seemed resize2fs was not capable of resizing a mounted filesystem in Redhat 4. In other words, the version of resize2fs that comes with the Redhat 4, can not enlarge a mounted filesystem.
While I was trying to find a way to make it work, the utility named ext2online helped me.
Actually it was the right utility to enlarge a mounted filesystem in Redhat 4 or in some of the other Linux Operating systems similar it.
[root@erpapp ~]# ext2online
ext2online v1.1.18 - 2001/03/18 for EXT2FS 0.5b
usage: ext2online [-C fd] [-dfqvV] device [new_size[bkmgt]]
-C, --completion : print completion information
-d, --debug : turn debug info on
-f, --force : skip safety checks
-q, --quiet : be quiet (print only errors)
-v, --verbose : be verbose
-V, --version : print version and exit
fd is the file descriptor to output completion data to,
new_size is in ext2 blocks (1k, 2k, or 4k) (default),
disk Blocks (512 byte), Kilo-, Mega-, Giga-, or Terabytes
[root@erpapp ~]# ext2online /dev/mapper/Cluster1-Root
ext2online v1.1.18 - 2001/03/18 for EXT2FS 0.5b
[root@erpapp ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/Cluster1-Root
450G 12G 416G 3% /
/dev/sda1 112M 17M 90M 16% /boot
none 16G 0 16G 0% /dev/shm
/dev/mapper/Cluster1-Temp
2.0G 288M 1.6G 16% /tmp
/dev/mapper/Cluster1-Log
So, at the end of the day, there is an important lesson learned. That is, when you are on Redhat 4, use ext2online to resize your mounted filesystems. (not the resize2fs).
Note: The rest of the approach was exactly the same as documented in http://ermanarslan.blogspot.fr/2013/10/linux-extend-logical-volume-size-lvm.html.
In that post, I gave all the info for resizing LVM based mount points. The explanation was from stratch. (scan scsi bus, fdisk, resize2fs and so on) and I have followed that approach many many times.
However, today, when I was working on a EBS application tier environment which was running on a RHEL 4 32 bit server, I noticed something and I want to share it with you.
I was requested to enlarge a mount which was based on LVM and which was mounted to root (/).
Again, I followed the same approach as documented in the blogpost, pointed by the above url but this was when I was enlarging the filesystem, the following error produced
[root@erpapp ~]# cat /etc/redhat-release
Red Hat Enterprise Linux AS release 4 (Nahant Update 5)
[root@erpapp ~]# resize2fs /dev/mapper/Cluster1-Root
resize2fs 1.35 (28-Feb-2004)
/dev/mapper/Cluster1-Root is mounted; can't resize a mounted filesystem!
So, it seemed resize2fs was not capable of resizing a mounted filesystem in Redhat 4. In other words, the version of resize2fs that comes with the Redhat 4, can not enlarge a mounted filesystem.
While I was trying to find a way to make it work, the utility named ext2online helped me.
Actually it was the right utility to enlarge a mounted filesystem in Redhat 4 or in some of the other Linux Operating systems similar it.
[root@erpapp ~]# ext2online
ext2online v1.1.18 - 2001/03/18 for EXT2FS 0.5b
usage: ext2online [-C fd] [-dfqvV] device [new_size[bkmgt]]
-C, --completion : print completion information
-d, --debug : turn debug info on
-f, --force : skip safety checks
-q, --quiet : be quiet (print only errors)
-v, --verbose : be verbose
-V, --version : print version and exit
fd is the file descriptor to output completion data to,
new_size is in ext2 blocks (1k, 2k, or 4k) (default),
disk Blocks (512 byte), Kilo-, Mega-, Giga-, or Terabytes
[root@erpapp ~]# ext2online /dev/mapper/Cluster1-Root
ext2online v1.1.18 - 2001/03/18 for EXT2FS 0.5b
[root@erpapp ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/Cluster1-Root
450G 12G 416G 3% /
/dev/sda1 112M 17M 90M 16% /boot
none 16G 0 16G 0% /dev/shm
/dev/mapper/Cluster1-Temp
2.0G 288M 1.6G 16% /tmp
/dev/mapper/Cluster1-Log
So, at the end of the day, there is an important lesson learned. That is, when you are on Redhat 4, use ext2online to resize your mounted filesystems. (not the resize2fs).
Note: The rest of the approach was exactly the same as documented in http://ermanarslan.blogspot.fr/2013/10/linux-extend-logical-volume-size-lvm.html.
Thanks!
ReplyDelete