UnixArena

Solaris ZFS – Interview Questions

By 11 Comments

As we know Oracle made ZFS is a default root filesystem from Solaris 11 onwards. This shows their future plans of improving ZFS and making it as worlds best volume manager cum filesystem in near future.Customers would like to have ZFS in their environment  since its a inbuilt volume manager and there is no additional operating cost unlike VXVM. They preferred ZFS data volumes in mid range servers and recommended as root FS in critical environment for its Liveupgrade facility. So each customer would like to have ZFS knowledgeable system admins on their environment.

Now a days you can expect more questions from ZFS in Solaris interviews. Here we will see some  of common ZFS interview questions. More ZFS tips can be found here.

Q 1.What is ZFS ? 
A. ZFS is filesystem cum volume managed developed by sun microsystems (Now its Oracle).It is a 128-bit file system.

Q 2.How to create new zpool ? 
A. The new zpool can be create using “zpool create pool_name cxtxdx cxtxdx ” 

Q 3.What type of raid levels are supported in ZFS ? 
A.

  1. Striped Vdev’s (RAID0)
  2. Mirrored Vdev’s (RAID1)
  3. Striped Mirrored Vdev’s (RAID10)
  4. RAIDZ (RAID5-Single parity)
  5. RAIDZ2 (RAID6-Double parity)
  6. RAIDZ3 (Triple Parity)
Q 4. Is it possible to reduce the zpool size ?
A. No .You can;t reduce the zpool size since all the disks will in stripped by default.

Q 5.How to list zpool’s ?
A. using “zpool list” command

Q 6. What is dataset in ZFS ? 
A. ZFS dataset is a virtual filesystem like volume in VXVM terminology.

Q 7.How to create a new dataset in ZFS ?
A.  Using zfs create pool_name/dataset_name

Q 8.How many LUNS or DISKS can be added to one ZPOOL ?
A. 264 LUNS or DISKS can be added to one zpool theoretically.

Q 9.How many dataset can be created per zpool ?
A.  264 dataset can be created in single zpool.

Q 10.How many zpools can be created in Solaris global zone ?
A. 264 zpools ca be created in a system.

Q 11.How to list the zfs snapshots ?
A. Use “zfs list |grep @ ” to get the list of snapshots.

Q 12. Is it possible to use fsck in ZFS ?
A. No.You can’t use fsck in ZFS.You can check the zpool integrity using “zpool scrub pool_name”

Q 13.Is possible to extend zfs dataset in online ?
A. Yes. you can increase the ZFS dataset size in online without un-mounting it .

Q 14.Is possible to extend zpool on fly ?
A. Yes. You can extend the zpool in online by just adding new disks using “zpool add pool_name cxtxdx ” command .

Q 15.What is minimum physical memory if you are using ZFS ?
A. Normally we recommend to use 4GB physical memory .

Q 16.Best practice of ZFS Mirrored Root Pool Disk Replacement.
A. You can perform the live disk replacement if the server supports hotplug and rpool is mirrored.
Step:1 
Offline the disk in zpool
# zpool offline rpool cxtxdx

Step:2 
Unconfigure the disk from OS level
# cfgadm -c unconfigure c1::dsk/cxtxdx

Step:3 
Physical replace the disk

Step:4 
Reconfigure the disk in OS level
# cfgadm -c configure c1::dsk/cxtxdx

Step:5 
Confirm that the replacement disk has an SMI label and a slice 0 to match the existing root pool configuration
Step:6
Making ZFS aware by using below command.
# zpool replace rpool cxtxdxs0

Step:6
Bring the disk in to zfs online.
# zpool online rpool cxtxdxs0

Step:7
Install the bootblock.
# installboot -F zfs /usr/platform/`uname -i`/lib/fs/zfs/bootblk /dev/rdsk/cxtxdxs0

Q 17. How to destroy zfs dataset ?
A. Using “zfs destroy pool_name/dataset_name”

Q 18.How to set new mountpoint to zfs dataset ?
A. Using “zfs set mountpoint=/new_mount_point pool_name/dataset_name
Q.17.How to rename ZFS dataset ?
A. We can use “zfs rename pool_name/dataset_name_old pool_name/dataset_name_new”

Q 18.Is it possible to recover destroyed zpool ?

Q 19.How to assign zpool/zfs dataset to local zones ?
Q 20.How to increase rpool size ?
A. We can increase the rpool size by using new big disks.

Q 21.Is it possible to convert root filesystem from UFS to ZFS ?
A. Yes. we can convert root filesystem from UFS to ZFS using liveupgrade .

Q 22.How to create ZFS dataset like vxvm/ufs volume ?
A. By setting quota and reservation for dataset ,we can get volumes like vxvm.

Q 23.Will snapshot consume any disk space when you create it ?
A. No.It will not consume any space when you create snapshot.The snapshot size will grow according to data changes in the volume.

Q 24. What is emulated volume in ZFS ?
A. It is a ZFS volume without ZFS filesystem on it.You can create UFS/VXFS on this.

Q 25.How to do the dry run on zpool?
A. You can use “-n” option to run dry run without performing any action.For an example if you want to know how the objects will be build using command without create the zpool,use -n option.
# zpool create -n arenpool raidz cxtxdx cxtxdx cxtxdx
would create 'arenapool' with the following layout:

arenapool
raidz1
    cxtxdx
    cxtxdx
    cxtxdx

Q 26. Is it possible to use ZFS as swap in solaris ?
A. Yes.We can use ZFS emulated  volume as swap.

Q 27.How to rename the zpool ? 
A. We can rename the zpool using below method.
Step:1  
Export the zpool
# zpool export pool_name

Step:2
Import the zpool name using new name.
# zpool import pool_name new_poolname

Q 30. Is it possible to create zpool using files ?
A. Yes. we can create a zpool using files which are created by using mkfile command. 
Ex:
# mkfile 64m /test
# zpool create arenzpool /test

Q 31.Do we need to update /etc/vfstab for ZFS datasets ?
A. By default,we no need to update vfstab for zfs dataset. If you set zfs mountpoint as legacy,then you have to update it .

Q 32.Do we need to maintain any configuration files for zpool ?
A. No.We no need to maintain any configuration files.By default all device based zpool will be imported and mounted according to the mountpoint value.


Q 33. How to perform the zpool scrub to check zpool integrity ?
A. use command “zpool scurb pool_name”

Q 34.Do we have to run mkfs command to create a ZFS on zfs datasets ?
A. No.We no need to create ZFS filesystem using mkfs.ZFS will automatically create filesystem and mountpoint when you create dataset. It will be automatically mount with dataset name if you didn’t mentioned the mountpoint there.

I would like every one to add ZFS related questions in the comment field.So that i can update this article with those questions and answers.

Thank you for reading this article. Good luck for interview.

Reader Interactions

Comments

    • Abdul says

      1. Volsize (Volume Size): Volsize refers to the maximum size that a ZFS dataset (such as a file system or volume) can grow to. It defines the upper limit of storage capacity that can be consumed by the dataset. When you create a ZFS dataset, you can specify its volsize to set this limit.

      2.Quota: Quota is a limit set on the amount of storage space that a specific user or group can consume within a ZFS dataset. It restricts the total storage capacity that can be used by files owned by the user or group. When a quota is set, users or groups cannot exceed the allocated storage limit within the dataset, regardless of the available space in the overall pool.

      3.Reservation: A reservation is a guarantee of a certain amount of storage capacity within a ZFS dataset. It ensures that a specified amount of space is reserved exclusively for a particular user, group, or application, regardless of the overall pool’s space availability. Reservations are typically used to prevent critical datasets from running out of space due to other datasets consuming all available space in the pool.

      Reply

    • Lingeswaran R says

      ZFS Volume – The Emulated volume is nothing but a portion of zpool space in a Raw format. You can even format this volume as UFS, Vxfs. You have to provide the size of the volume while creating it.

      ZFS dataset – the Soft lawyer of ZPOOL. When you set the Quota & Reservation to the dataset, it can act like a Volume with ZFS FS.

      Reply

    • Lingeswaran R says

      The refquota and refreservation properties are used to manage file system space without accounting for disk space consumed by descendents, such as snapshots and clones.

      quota & reservation will account snapshots & clones of the dataset.

      Lingesh

      Reply

      • ananda raman says

        @Lingeswaran..just for my clarify,lets consider we have zpool say newpool size 10gb and inside that we created dataset “test” with quota 8gb and reservation 5gb..Now i have created snapshot of the dataset ..in this scenario how refquota and ref-reservation will be useful .Please let me know

        Reply

Leave a Reply

Your email address will not be published. Required fields are marked *