FMS Backup Strategy Best Practices

[Elena2]

hello. I am extremely new to FMS. Is there a document anywhere that has best practices on server setup and Backup strategy in particular? Weekly full? daily? hourly; incremental backups? are there transactional backups? What are the issues, corruption issues? mistakes to avoid?

We need this urgently. Believe that the server was not initially set up properly and now, of course, I have inherited the issue  We are running FMS 14

any help would be appreciated,

[ggt667]

In the server control panel you will see Schedules; that's where you configure scheduled backups.

 

As for white paper I have not yet seen a good one.

 

Personally I prefer ZFS for storage as this allows for system snapshots, rollback, compression, and separate backups rooted in the snapshots. ZFS has a mac implementation named o3x.

Edited March 22, 2018 by ggt667

[Wim Decorte]

Your backup strategy needs to be set based on the answers to these two questions:

- how much data are you willing to lose?  This determines your 'restore points', how many backups you need to make

- how much down-time are you willing to incur?  This determines your 'restore time', your procedure for how quickly you need to be able to restore the files

The answers also determine obviously how much effort and money you'll need to invest in making this so.  

 

Incremental backups are as close to 'transactional' backups as we have in FM.  In FM they are called "progressive backups" and they keep 2 backup sets only.  Progressive backups are meant to be part of the overall backup strategy, and not to be the only backup used (mainly because it offers that limited number of restore points).

 

[Elena2]

thank you ggt667. I will take a look. but still hoping for documentation that I can point to as recommendations.

thank you Wim Decorte. I am going to get those 2 questions answered as soon as possible. But let's guess, this is marketing, so the answer will be none and 0. what would be the best strategy? Full daily, incremental hourly or more frequently? to a local drive. then multiple times a day move the backups to the SAN?

 

 

[Ocean West]

Good server deployment.

RAID for Live database (recommended not required)
RAID for backups (recommended not required) 
these are for uptime operations when drive does fail you can get thru the day and can replace drives at end of day, or swap with a spare drive ( but rebuilding array may degrade services - depending on how complex your solution and how many users you have)

Progressive Backups for rolling back faster to a good state to should a user or developer accidentally trashes some records - most likely this would be the least in terms of lost data. But does require that you are informed as close to the discovery as possible - your full backup may actually may have the desired restore state.

Full backups - Backing up hourly, weekly, monthly will require more storage to maintain files - and if you have many backups thru the work hours a second device on its own controller may be best for performance. 

Retention - This is depended on your company policy and preference - you may wish to go way back in time because someone did something and are just now coming forward where, you may need to restore layouts or data, or just want to take a trip down memory lane.

Offsite - you are not really backed up unless you have an offsite backup - You can use services / software that "backup a backup" to another offsite location - either controlled by you or a trusted backup company. - For third party backups you should really have all your files using Encryption at Rest (EAR)

Spot check - Drive failures, and data losses don't call you up and schedule an appointment they can and will happen - usually when you have other plans. However you can't just set up your backups and believe everything is working - the day you restore only to find out the file is corrupted or didn't backup a file - or remote containers etc. You need to be proactive and schedule at least a quarterly spot check where you pull backups from your different locations and open the files up and spend some time to review that your data and schema is indeed recoverable. 

 

[Elena2]

thank you Ocean West. I've also read all the articles for "backing up your data Overview" Knowledge Base article https://support.filemaker.com/s/answerview?language=en_US&anum=13793  and all the related links shown there.  I now know just enough to be dangerous!

[OlgerDiekstra]

I manage a Windows 2K8 R2 FM16 server. Full backups are done nightly (daily: Mo/Tu/We/Th, weekly on Fr, none on Sa/Su as no data gets changed then), with progressive enabled. C: drive is the OS drive and holds the Applications. D: drive is where the databases live. E: drive is where the backups are stored. C:, D: and E: are partitions on a RAID drive. A batch script copies backups nightly to a NAS. The NAS copies FM backups (along with other backups) to another NAS. The other NAS backs up to AWS. Everything is fully automated and requires no human intervention, aside from the checks. If backups fail, I get alerted. Generally, no news is good news, but the system is still regularly checked.

[Elena2]

Thank you OlgerDiekstra. That is a good plan. I appreciate all the replies I have received. All have excellent info for me. Thank you!