How To Set Up Bare Metal Recovery for Virtual Machines

Steve’s How-To Series

How To Set Up Bare Metal Recovery for Virtual Machines: (View this second installment) We all think of bare metal recovery for physical machines, but it is just as necessary for virtual ones. In this session, Technical Support Lead Steve Adler will demonstrate how simple it is up to set up bare metal recovery for virtual machines on the NetVault: Backup VMware Plugin. Watch the full video (5:31) on our BakBone YouTube channel.

How to Set Up Deduplication

Steve’s How-To Series
How to Set Up Deduplication: View this first installment of Technical Support Lead, Steve Adler’s, insight into maximizing your NetVault investment. He will demonstrate just how quick and easy it is to set up deduplication with NetVault: SmartDisk. Watch the full video (4:02) on our BakBone YouTube channel.

Quick Guide to NetVault: Backup Processes

Steve Adler, Technical Support & Operations Lead

There are a several different NetVault: Backup (NVBU) processes, and in this blog I’ll give you a quick overview of our different processes and what they do. Every NVBU process has its own unique job to do, and sometimes, that job requires more resources in order to complete the task. It’s only unusual if those resources aren’t released after the task has been completed, at which time, you should contact our Technical Support team. Sometimes, just knowing which tasks a process is performing will make it easier to understand why it uses the resources that it does.

nvsched: This is the NVBU Scheduler. It stores information about jobs in the Schedule Database, kicks off a job manager when it’s time for a job to run, and reschedules instances of repeating backups.

nvavp: The NVBU Auditor. The enforcer. The process that decides who can perform an operation in NVBU. The Auditor stores information about every user, oversees NVBU events and notifications, and verifies that when a user tries to do something within NVBU, they have the correct permission to do so. If they don’t have the authority, they are blocked from performing that action.

nvjobmgr: This is the one that throws people off, especially when they have multiple jobs queued up to run. The NVBU Job Manager is invoked by the Scheduler when the time arrives for a job to run. You will have one instance of the nvjobmgr process per job and it will run until that job has completed. It manages the job state, (running, waiting for media, etc..). It also coordinates with all the other processes so that they work as a team to get the job done.

nvfs: The NVBU FileSystem Plugin. Quite simply put, this is the process that allows you to browse, backup from, or restore to a machines filesystem. This will run on any machine that is the target of a filesystem job.

nvmedmgr: NVBU Media Manager. Quite possibly the hardest working process. The Media Manager process stores all the information about every piece of media known to the NVBU Server. It handles all requests for every piece of media required for a job.

nvdevmgr: NVBU Device Manager. These processes communicate with the tape drives, SmartDisk, etc.. Allows us to read from, write to, position, load, and unload media as required to complete a job. You will have one process per drive.

Duplicating Backups – Selecting between Duplication and Data Copy

Steve Adler, Technical Support & Operations Lead

We back up our data so that in case of a data loss event, we have the ability to restore that data. But how do you protect the data that you’ve backed up. You can send it offsite, but what happens if that offsite location is affected by a catastrophic event? It’s simple really. Make another copy!

NetVault: Backup provides two different ways to make additional copies of your backups. Duplication and Data Copy. What’s the difference?

The Duplication method creates an exact secondary copy which is linked to the original primary backup. During duplication, the copy is broken into segments and stored on the secondary backup device. During restore, segments from the primary copy and the secondary copy are interchangeable. This means that even if you no longer have all of the original tapes, as long as the media that contains the duplicates of that data is available, you’ll be able to restore that data. You can basically “fill in the blanks” with the duplicated data.

However, encrypting that data for added security can pose a problem. It is impossible to unencrypt the primary backup and encrypt the secondary copy because it is not possible to mix unencrypted segments with encrypted segments during restore. If the original saveset is encrypted, the Duplicate method will create an encrypted secondary copy. If you have no encryption for the primary backup, the secondary copy will also be unencrypted.

The Data Copy method is recommended when you want to create a secondary copy for offsite storage. Data Copying a backup breaks the backup into segments and copies the segments onto the targeted backup device. During restore, NetVault: Backup only restores the primary copy or the secondary copy. Backup segments from the primary and the secondary are not interchangeable. This means that if you are missing part of the original backup, you will need the entire secondary copy available to perform a restore. You can’t just “fill in the blanks.” However, this enables the ability to encrypt the Data Copy, or secondary copy, while the primary copy remains unencrypted such as with deduplicated primary backups.

By incorporating one or both of these strategies to copy your data, you can decrease the chance of losing your backups to a catastrophic event so you can recover your data when you need it.

What is Shared Memory and How Does it Affect NetVault: Backup?

Steve Adler Technical Support Lead

 

First, we need to know what shared memory is…

Shared memory is a method of inter-process communication (IPC), where two processes communicate with each other through shared blocks of RAM. Because communication is resident in RAM, shared memory may be simultaneously accessed by multiple programs and will either provide communication among them or avoid redundant copies. Modifying shared memory is typically done on performance oriented systems, such as databases or backup systems.

In Windows and Solaris, shared memory is handled dynamically.  In Linux and OS X, it’s not.

In Linux and  OS X, these settings need to be changed if you are will be running a lot of NetVault: Backup Jobs at one time, or backing up massive amounts of data.

So what is shared memory made up of?  How is it defined?

kern.sysv.shmall
shmall represents the maximum number of pages able to be provisioned for shared memory. It determines the total amount of shared memory that the system can allocate. To determine total system shared memory, multiply this value by the size of the page file. The page file size can be determined via `vm_stat` or `getconf PAGE_SIZE`. A typical page size is 4KB, 4096 bytes. In OS X, Apple uses extremely conservative settings for shmall. At 1024,  OS X defaults to only 4MB of shared memory which isn’t going to be enough for your backup environment

kern.sysv.shmseg
shmseg represents the maximum number of shared memory segments each process can attach. Default in OS X is 8.

kern.sysv.shmmni
shmmni limits the number of shared memory segments across the system, representing the total number of shared memory segments. Default in OS X is 32.

kern.sysv.shmmin
shmmin is the minimum size of a shared memory segment. The default is one, and there is no reason to change it.

kern.sysv.shmmax
shmmax is the maximum size of a segment. Default in OS X is 4 MB.  This is going to be your performance bottleneck.

Linux defaults vary for each distribution.

How does this affect NetVault: Backup?

NetVault makes extensive use of shared memory for it’s own inter-process communication.  If you modify the Process Manager and NetVault fails to start, it’s probably because you’ve told NetVault to use more shared memory than is actually being provided by the OS.  So, we’ll need to make the OS share more.

In my experience, you can safely set your shared memory to half of your physical RAM without any performance degredation.  However, if you have 4+ gigs of RAM, it’s not necessary unless you have a database that is really, really active.

How do you do this?

In OS X 10.5 and above, you create a file called /etc/sysctl.conf and then add the lines in the example below and reboot.  In 10.4 and below, you would create a file called /etc/rc and do the same.

In Linux, you will modify the /etc/sysctl.conf file just like OS X 10.5.

Examples:

512MB of shared memory
kern.sysv.shmall: 131072
kern.sysv.shmseg: 32
kern.sysv.shmmni: 128
kern.sysv.shmmin: 1
kern.sysv.shmmax: 536870912

1GB Shared memory
kern.sysv.shmall: 262144
kern.sysv.shmseg: 32
kern.sysv.shmmni: 128
kern.sysv.shmmin: 1
kern.sysv.shmmax: 1073741824

Note the relationship between shmall and shmmax. shmall x 4096 = shmmax.

Hopefully this will help you get a little more performance out of your Mac or Linux machines.