StorageManagement.md

Storage Management

List, create, delete, and modify physical storage partitions

• lsblk lists all available disk devices plus available partitions

• fdisk it is used to manage disk partition in MBR modality

  • E.g. fdisk /dev/sda

    This will open an interactive menu that will permit to show current status of partitions or create a delete new partitions

• gdisk it is used to manage disk partition in GPT modality

  • E.g. gdisk /dev/sda

• Destroy all MBR partition on a disk

  • gdisk /dev/sda -> x (expert) -> z (zap)

• Convert MBR to GPT

  • gdisk /dev/sda -> W -> Y

Manage and configure LVM storage

• Before create a Logical Volume must be created in sequence a physical volume and after a volume group
• A physical volume is a partition that can be part of volume group. Inside a volume group can be created logical volume
• The advance of logical volume is that their dimension can be managed easly
• If more space is need a volume group can be extended as well

Physical Volume

• pvcreate /dev/sdb1

  To create a physical volume with partition sbd1

• pvs lists available physical volumes

• pvdisplay /dev/sdb1 shows info of a physical volume

Volume Group

• vgcreate vgname /dev/sdb1

  To create a volume group called vgname and add the sdb1 physical volume to it

• vgs lists available volume groups

• vgdisplay vgname shows info of a volume group

• vgextend vgname /dev/sdc3 extends a volume group adding a new physical volume /dev/sdc3

Logical volume

• lvcreate -n volumename -L 10G vgname

  To create a logical volume called volumename of size 10GB on volume group vgname

• lvcreate -n volumename -l 100%FREE vgname

  To create a logical volume called volumename with all available space on volume group vgname

• lvs list available logical volumes

• lvdisplay shows info of all logical volumes

• lvdisplay vgname/volumename shows info of a logical volume volumename contained in vgname volume group

• Before use a logical volume, a file system must be created on it

• blkid /dev/vgname/volumename shows the UUID of a formatted volume group

• lvextend -L +1G -r vgname/volumename extends the logical volume volumename of one giga

  • -r is used to resize file system

• lvreduce -L -1G -r vgname/volumename reduce the logical volume volumename of one giga

Create and configure encrypted storage

• To use encrypted storage a kernel module must be loaded
  • sudo modprobe dm_crypt Loads kernel module dm_crypt
  • echo dm_crypt >> /etc/modules-load.d/dm_crypt.conf to load dm_crypt module automatically when system will be restarted
  • lsmod lists all loaded kernel modules
• yum -y install cryptsetup install software used to manage encrypted storage

Encrypt

• cryptsetup luksFormat /dev/vgname/volumename encrypts a logical volume volumename contained in vgname volume group

  • A password must be provided

  • When confirmation will be required insert a capital YES

  • NOTE: this command can be used with physical volume as well

• cryptsetup open --type luks /dev/vgname/volumename namenewdevice

  It open encrypted volume and associate it to a new device called namenewdevice

  • Password must be provided

• mkfs.ext4 /dev/mapper/namenewdevice

  It creates a file system in namenewdevice

  Now new the new device can be mounted

Close device

• Unmount device
• cryptsetup close namenewdeviceclose namenewdevice

Automount

• echo "passwd" >> /root/key Insert a string that will be used that will be used as authentication key to open device

• chmod 400 /root/key reduces permission on key file

• cryptsetup luksAddKey /dev/mapper/namenewdevice /root/key add key to encrypted device called namenewdevice

• Edit /etc/crypttab and add below row:

  • namenewdevice /dev/vgname/volumename /root/key

• Add below row to /etc/fstab

  • /dev/mapper/namenewdevice /mnt/mountpoint ext4 defaults 0 0

• Reboot system or reload system manager

  • systemctl daemon-reload

• The new encrypted volume will be mounted on /mnt/mountpoint

Configure systems to mount file systems at or during boot

• Edit /etc/fstab adding a row similar to:

  • /dev/sdb1 /mnt/mountpoint ext4 defaults 0 0

    • Mount device sdb1 to mountpoint.

    • Device is formatted using ext4 filesystem.

    • Default mount options are used

    • 0 0 -> Dump (bkp) and fsck.

      • First 0 means no backup required
      • Second 0 means no fsck required in case of not correct umount. To enable fsck insert 2 because number indicate the check order, and 1 is given to operating system disk and two do data disks

• mount shows mounted volumes

• mount -a reloads /etc/fstab

• mount -t type -o options device dir

  • It mounts a device formatted with file system type on directory dir using a list of options

  • options can be:

    • async -> I/O asincrono
    • auto -> Can be mounted using mount -a
    • default ->Equal to this list of options: async,auto,dev,exec,nouser,rw,suid
    • loop -> To mount an ISO image
    • noexec -> no exec
    • nouser -> A user cannot mount this volume
    • remount -> Mount volume also if it is already mounted
    • ro -> Read only
    • rw -> Read an write
    • relatime -> Modify file access time (atime) if file is changed or one time a day. Alternative, to reduce disk traffic, noatime can be used. This is useful with SSD to avoid not useful write.

SMB protocol

• yum -y install samba-client cifs-utils it installs software need to manage CIFS/SMB protocol

• smbclient -L targetIP

  It lists all SMB shared directory available on a target IP

  • root password must be provided

• mount -t cifs -o username=smbuser,password=1234pwd //192.168.0.10/share /media/samba

  It mounts a directory share, shared by server 192.168.0.10 on samba directory. User and password to authentication are provided

• Permanent configuration

  • echo "username=smbuser" >> /media/smb/.smbconf
  • echo "password=1234pwd" >> /media/smb/.smbconf
  • chmod 600 /media/smb/.smbconf
  • In /etc/fstab insert:
    • //192.168.0.10/share /media/samba cifs credentials=/media/samba/.smbcredentials,defaults 0 0

NFS protocol

• yum -y install nfs-utils it install software to manage NFS protocol

• showmount -e targetIP

  It lists all NFS shared directory available on a target IP

• mount -t nfs -o defaults 192.168.0.10:/srv/nfs /media/nfs

  It mounts a directory nfs, shared by server 192.168.0.10 on nfs directory

• Permanent configuration

  • In /etc/fstab insert:
    • 192.168.0.10:/srv/nfs /media/nfs nfs defaults 0 0
  • To user NFSv3 insert:
    • 192.168.0.10:/srv/nfs /media/nfs nfs defaults,vers=3 0 0

Configure and manage swap space

• To use a device as swap space:
  • mkswap /dev/sdb3
  • swapon -v /deb/sdb3
  • In /etc/fstab insert:
  • • /dev/sdb3 swap swap defaults 0 0

Create and manage RAID devices

Concepts:

• Parity disk. It is used to provide fault tolerance.

• The spare device. It not take part of RAID and it is used only in case of a disk fault. In this case spare enter in the RAID and the content of lost disk is reconstructed and saved on it.

• yum -y install mdadm installs software to manage RAID devices

• RAID 0 - Striped - No spare

  • mdadm --create --verbose /dev/md0 --level=stripe --raid-devices=2 /dev/sdb1 /dev/sdc1

• RAID 1 - Mirror

  • mdadm --create --verbose /dev/md0 --level=1 --raid-devices=2 /dev/sdb1 /dev/sdc1

• RAID 5 - (1 parity + 1 spare)

  • mdadm --create --verbose /dev/md0 --level=5 --raid-devices=3 /dev/sdb1 /dev/sdc1 /dev/sdd1 --spare-devices=1 /dev/sde1

• RAID 6 - (2 parity + 1 spare)

  • mdadm --create --verbose /dev/md0 --level=6 --raid-devices=4 /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sde --spare-devices=1 /dev/sdf1

• RAID 10 - (Stripe + Mirror + 1 spare)

  • mdadm --create --verbose /dev/md0 --level=10 --raid-devices=4 /dev/sd[b-e]1 --spare-devices=1 /dev/sdf1

• mdadm --detail /dev/md0 shows status of RAID device

• To use device md0, format it and use as a classical device

Monitoring RAID devices

• mdadm --assemble --scan
• mdadm --detail --scan >> /etc/mdadm.conf
• echo "MAILADDR root" >> /etc/mdadm.conf
• systemctl start mdmonitor
• systemctl enable mdmonitor

Add disk

• mdadm /dev/md0 --add /dev/sbc2

• mdadm --grow --raid-devices=4 /dev/md0

  It adds a spare disk and after it grows array

Remove disk

• mdadm /dev/md0 --fail /dev/sdc1 --remove /dev/sdc1

  mdadm --grow /dev/md0 --raid-devices=2

  It mark disk as failed and remove it. After the size of array must be adjusted

Delete RAID

• Unmount device
• mdadm --stop /dev/md0
• mdadm --zero-superblock /dev/sbc2 It clean partition that, after, can be reused

References:

• https://raid.wiki.kernel.org/index.php/A_guide_to_mdadm

Configure systems to mount file systems on demand

• yum -y install autofs installs software need to manage automount

Automount NFS directory

• Edit /etc/auto.master and insert:

  • /media /etc/nfs.misc --timeout=60

• Edit /etc/nfs.misc and insert:

  • nfs -fstype=nfs 192.168.0.10:/srv/nfs

• systemctl start autofs

Create, manage and diagnose advanced file system permissions

ACL Access control list

• They must be supported by filesystem

• With some old filesystem a mount option (e.g. acl) must be provided to enable ACL

• getfacl file shows ACL applied to a file

• setfacl -R -m g:sales:rx file set ACL on file

  • -R recursive, if file is a directory, ACL will be applied to all file inside it
  • -m modify
  • g:sales:rx group sales can read and execute
    • g group
    • u user
    • o other

• setfacl -m u:dummy:- file remove all permissions of user dummy.

• setfacl -m d:g:sales:rx directory set a default ACL to a directory. In this way all files created inside it will have same ACL as default

  The default ACL is a specific type of permission assigned to a directory, that doesn’t change the permissions of the directory itself, but makes so that specified ACLs are set by default on all the files created inside of it

• If an ACL is applied, when ls -la is executed an + is inserted after other permissions

• setfacl -x u:test:w test remove ACL

• setfacl -b file removes all ACL

Extended attributes

• They are file properties

• With some old filesystem a mount option (e.g. user_xattr) must be provided to enable extended attributes

• Only root user can remove an attribute

• chattr +i file add immutable attribute to a file. It cannot be deleted or removed

• chattr -i file remove immutable attribute from a file.

• lsattr file shows file's extended attributes

Setup user and group disk quotas for filesystems

• Quota: space that can be used by an user on one specific filesystem
  • NOTE: To limit space in a directory it is better create a specific mount point with a specific partition
• yum -y install quota installs software need to manage quota
• usrquota,grpquota mount options must be inserted for filesystem to which enable quota (e.g. editing /etc/fstab)
• After that options are inserted, remount partition to enable them
• After remount execute quotacheck -mavug that check used blocks and inserted them in a tracking file
  • Two files will be created:
    • aquota.group
    • aquota.user
• quotaon -a start quota system
  • Alternative:
  • quotaon -vu /mnt/mountpoint it starts only quota user for specific mountpoint
  • quotaon -vg /mnt/mountpoint it starts only quota group for specific mountpoint
• quota -vu user shows user's quota
• The quota is specified in blocks of 1K size and in number of inode that is the number of files that can be created
• Hard limit: maxim value allowed
• Soft limit: a limit that can be exceeded for a grace period. Default grace period is a week
• When grace period is reached, soft limit become and hard limit
• edquota -t Edit the grace period. Is an unique value for all system
• edquota -u user edit user's quota
  • In each column can be insert a value for soft and hard limit for blocks and inode
  • NOTE: Normally soft and hard limits are configured equal to avoid confusion
• repquota -aug It shows an overview of current quota for each users

Create and configure file systems

• mkfs.ext4 /dev/sdb1 creates an filesystem ext4 on sdb1 partition
• fsck.ext4 /dev/sdb1 checks the integrity of sdb1 filesystem