Mount and unmount CIFS and NFS network file systems

Let's start with NFS first.
Requirement number one is to have an NFS server available and configured. As part of the RHCE objectives you need to configure one.

The mount command for nfs shares is straight forward enough:

mount.nfs 10.168.20.227:/distro /repo - where 10.168.20.227 is an NFS server that has exported a directory called /distro to, at the very least, your computer, more likely it will be to your entire network and you want it to be mounted to /repo. Note that you can simply use mount, instead of mount.nfs.

Let's have a look now at CIFS shares. Again, you need to have CIFS shares available.

mount.cifs "//10.168.20.112/c$/" test -o user=roots dom=dev - this is the equivalent to \\10.168.20.112\c$ in windows notation, with the user roots in the domain dev. Note that you can pass the password as a parameter too, simply add to the end of the command pass=rootspassword, if you don't you'll get a prompt to enter the password.

Note that you need to install nfs-utils (yum install nfs-utils -y) to mount nfs shares and cifs-utils (yum install cifs-utils -y) to mount cifs shares.

Mount, unmount and use LUKS-encrypted file systems

Shameless plug this one. It is really covered by the last post for information on how to mount partitions and this post for creation of LUKS-encrypted file systems.

Create, mount, unmount and use ext2, ext3 and ext4 file systems

In order to create a file system, you need to use the command mkfs. Depending on filesystem type you can use:

mkfs.ext2 <devicename>
mkfs.ext3 <devicename>
mkfs.ext4 <devicename> 

Thus for instance, in order to create a ext4 filesystem with label MyLabel you would run the following command:

mkfs.ext4 /dev/sdc1 -L MyLabel

In order to mount this newly created filesystem to a directory called MyLabel, you would type:

mount /dev/sdc1 /MyLabel
or
mount -L MyLabel /MyLabel

Unmounting is done by the following command:

umount /MyLabel

Note that all of this mounting is temporary, so if you want to make it permanent, then you need to add it to the /etc/fstab file, e.g.

echo "LABEL=MyLabel /MyLabel ext4 defaults 1 2" >> /etc/fstab 

Add new partitions, logical volumes and swap to a system non-destructively

This objective is really more of the same as this and this. The only new part is the adding of swap space.
Again, this can be done with fdisk and parted. Let's have a look at fdisk first.

Assuming that you have device available on /dev/sdc and you want to use it all

1.  fdisk /dev/sdc
2. Type p
3. Type 1
4.  Hit enter twice
5. Type t
6. Type 82
7. Type w

We now need to make it a swap partition


mkswap /dev/sdc1


and now we can activate with


swapon /dev/sdc1


You can check it with


swapon -s

Filename                                Type            Size    Used    Priority
/dev/sda2                               partition       2047992 0       -1
/dev/sdc1                               partition       1044184 0       -2

Now we need to add it to the /etc/fstab file, which can be done with the following command:


echo "/dev/sdc1 swap swap defaults 0 0" >> /etc/fstab

Configure systems to mount file systems at boot by Universally Unique ID (UUID) or label

A nice and short objective.

Remember in my last post how I added a label to the my luks encrypted logical volume, I gave it the label LUKS, so I can now mount it with the following command:

mount -L LUKS /encdir

The UUID can be found by issuing the following command:

blkid - This has this edited output

/dev/mapper/volname-lvol3: UUID="d1e50970-2995-4749-b9b1-46a52169a305" TYPE="crypto_LUKS"
/dev/mapper/lukslv: LABEL="LUKS" UUID="d71b791d-3f09-4419-bedd-ae9bc0ed0989" TYPE="ext4"

The first device represents the logical volume itself, whereas the second device is what we are interested in, the encrypted device, thus the UUID we want is d71b791d-3f09-4419-bedd-ae9bc0ed0989.

So, we can also mount by using this command:

mount -U d71b791d-3f09-4419-bedd-ae9bc0ed0989


Of course, the two commands above will only temporary mount the device, so we need to edit the fstab file again.

To mount by label you can do it with the following command:

echo "LABEL=LUKS /encdir ext4 defaults 1 2" >> /etc/fstab

and to mount by UUID you need to modify the fstab file like this:

echo "UUID=d71b791d-3f09-4419-bedd-ae9bc0ed0989 /encdir ext4 defaults 1 2" >> /etc/fstab

You could also edit /etc/fstab. At any rate, make sure that you only have one entry that mounts the encrypted device.

You can test your /etc/fstab file by issuing the following command, which will attempt to mount everything on /etc/fstab: mount -a

Create and configure LUKS-encrypted partitions and logical volumes to prompt for password and mount a decrypted file system at boot

In the past two posts, I have explained how to create a primary or logical partition or a logical volume, so you'll need to create one of them and fill it with random data for security reasons, note that this could take quite a while to complete. Since I already have a few logical volumes available, I'll use them. This has the upside of not taking very long as they are fairly small. Let's fill it with random data then:

dd if=/dev/urandom of=/dev/volname/lvol3 bs=1M

Once this is completed, you can create your LUKS partition with the following command:

cryptsetup luksFormat /dev/volname/lvol3 -- you'll get the following output. Just follow the instructions

WARNING!
========
This will overwrite data on /dev/volname/lvol3 irrevocably.

Are you sure? (Type uppercase yes): YES
WARNING!
========
This will overwrite data on /dev/volname/lvol3 irrevocably.
Enter LUKS passphrase:
Verify passphrase:

Make sure that you remember your passphrase, you'll need it soon enough. Now, create the mapping to the encrypted device, logical volume in this case, by issuing the following command:

cryptsetup -vv luksOpen /dev/volname/lvol3 lukslv

You now have an encrypted logical volume called lukslv.

Let's give this badboy a file system:

mkfs.ext4 /dev/mapper/lukslv -L LUKS

Note how I have give it a label (-L LUKS), this is not necessary but can be helpful.

Next stop, /etc/crypttab.

echo "lukslv /dev/volname/lvol3 none" >> /etc/crypttab

will set up this encrypted logical volume at boot time.

You can now add the device to the /etc/fstab file so that it is mounted on boot up.

echo "/dev/mapper/lukslv /encdir ext4 defaults 1 2" >> /etc/fstab

I have created a directory called /encdir to test this.

If you still remember your passphrase, you can reboot now.

init 6

You'll get this prompt when booting up, just enter your passphrase and you are good to go:

Create and remove physical volumes, assign physical volumes to volume groups, create and delete logical volumes

This is another objective that is not very useful by itself, after all once you create a logical volume, you still have to create a file system and mount it, anyway, let's get cracking.

Assuming you have an empty disk you can simply do the following to create a physical volume:

pvcreate /dev/sdc - this assumes that you have a a device called /dev/sdc


This is not very useful, namely to have a single physical volume, the only thing you have done is given yourself the oportunity to add more storage later, which is not actually a great thing, but for this objective you are better off with at least two disks.

Before we move any further, let's cover the remove physical volumes objective:

pvremove /dev/sdc

There are limitations to what the command can do, e.g. don't expect it to remove a physical volume that has a volume group.

Note that the -ff option will remove it almost no matter what, use at your own risk.

You can list the physical volumes with the following commands:

pvs

pvdisplay

Assuming that you have created two physical volumes, /dev/sdc & /dev/sdd, we can create a volume group like this

vgcreate volname /dev/sdc /dev/sdd

Alternatively, you could this to mimic a real life situation where you create a logical volume, run out of space and need to extend, in which case you add a new physical volume to the volume group:

vgcreate volname /dev/sdc - this creates the volume group volname

vgextend volname /dev/sdd - this adds the /dev/sdd phyical volume to the volname volume group

Say you made a mistake and you actually wanted to add /dev/sde rather than /dev/sdd to volname, just run the following

vgreduce volname /dev/sdd - needless to say that there will be limitations to this command as well

You can list the volume groups with the following commands:

vgs

vgdisplay

And finally we come to the useful part, the logical volumes, by now, you probably have sussed out what the commands are going to be: lvcreate, lvremove, lvs, lvdisplay, 100% right.

lvcreate -l 10%VG volname -n lvol1 - This will create a logical volume called lvol1 that is 10% the size of the volumegroup volname.

lvcreate -L 500M volname -n lvol2 - This will create a logical volume called lvol2 that is 500M, this will only work if there is enough free space in volname.


lvcreate -i 2 -l 10%FREE volname -n lvol3 - This will create a logical volume called lvol3 that is 10% of  the size of the space left in volname. The -i 2 switch will ensure that volume group spans two physical volumes, again this will only work if there are two or more physical volumes.

You've just realized that you wanted a bigger lvol3, no problem, just extend it with:

lvextend -l 20%FREE /dev/volname/lvol3

The lvreduce command behaves in a similar manner

In order to delete the lvol3 you would type:

lvremove /dev/volname/lvol3
or
lvremove /dev/mapper/volname-lvol3

List, create, delete and set partition type for primary, extended, and logical partitions

As with many things in Linux, there are many ways of performing the tasks needed to master this objective. In general you will either use fdisk or parted. 

List Partitions: 

fdisk -l - will provide a list of partitions for all devices in the system

fdisk -l /dev/sda - will list all partitions for disk sda

Alternatively, you can use parted, like this:

parted - once inside the parted shell, type:

print all

In order to list the partitions in a single device, once inside the parted shell, simply type:

print

If you want to change the device you are operating in, you can either invoke add the device to the parted command, e.g.

parted /dev/sdb

or while inside the parted shell:

select /dev/sdb

Note, that you could run the commands outside the parted command shell, like this: parted -l

Create Partitions:

First we'll use fdisk, note that this is for a blank disk, if the disk already contains partitions then there will be limitations, e.g. if you already have 4 primary partitions you'll only be able to create an extended partition.

1. Start fdisk (note that you can type m for help) - fdisk /dev/sdb  or fdisk -u /dev/sdb for sizes in sectors instead of cylinders
   
2. Type n and you'll get the following options

   Command action    

e   extended   

         p   primary partition (1-4)  

3.  Type p
4. Type 1, to make this the first partition
5. The next option is where to start your partition, I normally accept the default, so I hit Enter
6. The last option is where to end your partition. If you want just one partition accept the default, otherwise just select your partition size, e.g. say you want a 1 GB partition, you need type +1G
7. type w, to write your changes to the partition table and exit fdisk.

You'll have to wait until the following objective  Create, mount, unmount and use ext2, ext3 and ext4 file systems to learn how to use the partition you have just created.

Now, let's turn to parted.

1. Start parted (note that you can type help for help)- parted /dev/sdb
2. Type makepart
3. You'll get a choice of primary or extended, type primary
4. Select the file type, e.g. ext4
   
5. Select the start for the partition, note that unlike fdisk, you have to type something here, e.g. 1 for the partition to start at the beginning of the disk
   
6. Select the end of the partition
   
7. You might get a warning that the partition is not best aligned for performance, if you set the start at 0.
8. Exit parted.
9. Alternatively, you can just type parted /dev/sdb primary ext4 1 1000M from the command shell
   

Note, that due to limitations in the MBR, it is only possible to have four primary partitions. If more partitions are needed, then an extended and logical partition(s) need to be created. In essence, an extended partition is created and inside the extended partition, logical partitions can be created. These can then be used in a similar way to primary partitions.

Using parted only, and assuming that the disk is 8 GB and has a 1 GB primary partition:

1. First make an extended partition - parted /dev/sdb mkpart extended 1000M 8000M
2. Then make a logical partition - parted /dev/sdb mkpart logical 1001M 4000M
3. output from: parted /dev/sdb print

Model: VMware Virtual disk (scsi)
Disk /dev/sdb: 8590MB
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start   End     Size    Type      File system  Flags
 1      1049kB  1000MB  999MB   primary   ext4
 2      1000MB  8000MB  6999MB  extended               lba
 5      1001MB  4000MB  2999MB  logical

Note how the logical partition has a number of 5

Delete partitions:

For fdisk.

1. fdisk /dev/sdb
2. Type d
3. Type the number of the partition you want to delete
4. Type w

For parted.

1.  parted /dev/sdb
2. rm n - where n is the number of the partition you want to delete.
3. Alternatively, parted /dev/sdb rm n - where n is the number of the partition you want to delete.

Set partition type:

This is relatively simple for fdisk.

1. fdisk /dev/sdb
2. Type t
3. Type the partition number you want to use
4. If you know the code, then type it, otherwise type L to get a list of partition codes
5. Let's say we want to make it a swap partition, so type 82
6. Type w
   

I'm not 100% sure how to do this with parted.

You can do parted /dev/sdb 5 set lvm on, to set the lvm label on, but I don't know how to change the partition type to say VMWare VMFS.

Start, stop and check the status of network services

Another relatively simple objective.

Start:

service servicename start

/etc/init.d/servicename start

Stop:

service servicename stop

/etc/init.d/servicename stop

Check Status


service servicename status

/etc/init.d/servicename status

Some services will have different commands available, just type service servicename to see them, e.g.

service sshd
Usage: /etc/init.d/sshd {start|stop|restart|reload|force-reload|condrestart|try-restart|status}

service httpd
Usage: httpd {start|stop|restart|condrestart|try-restart|force-reload|reload|status|fullstatus|graceful|help|configtest}

Start and stop virtual machines

A short and sweet objective this one:

Start Virtual Machines:

virsh start domain

Stop Virtual Machines:

virsh shutdown domain - This initiates a shutdown, so it may take a while for the machine to go down.


virsh destroy domain - This is equivalent to disconnecting the electricity supply from the server and thus it is immediate.