That's basically how I have my 2 truenas boxes setup. Been running them for years now. It works perfectly 😊 That said I still use core. I'll go to scale when core goes EOL
For encryption, the fact you must type a directory name to make a new directory instead of making a dataset and choosing it, is extremely confusing and impossible without this video. Thank you so much. for more info, no matter the property you set for the designation it will not work, unless you double encrypt (encrypted designation inherited folder, and re-encrypt the snapshots sent to put in there).
so niceeee. Just what I needed now as I bumbed into problems. Got a secondhand HP DL380 Gen9 to get rid of my very loud and very old IBM 360. Hope I can manage the replication with this tutorial. also boot disk is failing on the old system and dont know if I can replace that so I move to the new system.
Great video. I'm just about to build a 2nd TrueNAS system to do just his. One question: I plan on replicating from the current system that is still running FreeNAS-11.3-U5 to a new system running the latest version of TrueNAS SCALE. Do you think this will be a problem? What can I do instead? Thanks.
Thank you so much for this. It helped me realize what I need to do for backup. Question I have , does the drive capacity of the backup server need to be as large as the data set that is snapshot? I ask because say all the data on my main pool is gone. 4 drives dead, z3, so the data is essentially gone. Is it possible to use these snapshots to layer back onto a new pool of drives say larger in capacity than before? That are currently blank? Or would those existing drives also need to be used? Would it be possible to recreate from scratch like that? I’m scared to death to lose my zpool disks and need someway to replicate the data if kind of a total failure happens
@LAWRENCESYSTEMS A question that I see myself and others struggle with about snapshots and restoring as a backup solution is: If I just start creating snapshots on an already established dataset that is 4TB in size, and a year from now I need to rebuild that dataset from a catastrophic failure, how would it be rebuilt from the most recent snapshot if it only reflects the last changes of say, a single pdf that was added? Wouldn't restoring with this snapshot only restore what it had detected as what changed since the previous snapshot and nothing before that? I have a hard time wrapping my head around how a 10kb snapshot can restore a 4TB+ dataset with this.
This definitely seems to be the Ticket for Full speed no overhead Truenas-Scale backups/Replications for DATA. I also like you can turn the backup/Replication Truenas Box into a Big JBOD Mixed Capacity Disk so I was able to put together a 35TB backup repository easily with some Misc Large/Med TB Drives i had to just serve as a backup form my ZFS Raided data. Even Still it replicated at 2.5GB/s on HDDS (Not SSDS) maxing out the 2.5GB network in my home. I tried to do this with Veeam B&R CE via - W10/11 Backing up Proxmox VM and it was stuck Sub 1GBPS for what ever reason Didn't mater if i run the Veeam B&R CE on another HEDT or a worker on the destination off the Backup Box or with a Root install of windows just slow never broke 1GBPS I ended up giving up i just have way to much TB for it to not get a full 2.5.
Thank you for the great tutorial. I successfully replicated my dataset, but I'm having trouble replicating "ix-application." I'm running Immich (App) on my primary TrueNAS Scale server. How can I replicate it to my offline TrueNAS server?
I would like to replicate from Core to Scale. Everything is setup propperly with SSH. Problem is after just a few moments I am getting this error and I don't know which permission are wrong and how to fix that. :( Replication "Datasetname" failed: cannot receive new filesystem stream: permission denied warning: cannot send "Snapshotname" signal received.
I've seen a lot of videos like this in setting up replication, but I haven't found anything on restoring TrueNAS from a replication after a complete failure. If system A replicates to system B, how do I restore system A from B after system A fails?
Okay, I had a misunderstanding of replication. When a task is completed the files are written to the other pools and appear in a filesystem the same way they are on the source end so you can interact with them. I thought replication stored it in a way where the data wasn't accessible unless recovered. I'll be using this for sure! I have two TrueNAS scale systems one at my house, one at my parents, both running Plex. Would like to back up there to their Plex library so they can also view the movies I add and it seems like this would be a good solution. Does that sound about right?
Thanks for the video. Replication from scratch is still a bit unclear to me. Let's say the backup server has replication data from multiple TrueNAS servers. Will Replication from scratch delete all snapshots regardless of destination paths, or just the snapshots for a set destination path?
If you manually snapshot the source and the destination, can it determine differences and copy just those or will the destination allways be wiped out? this was not clear.
You can have it send over all snapshots from source both manual and automatic, but don't make changes on the destination because they will either be over written or cause the process to fail.
Perfect, just what I need! So, the replication task always does a RAW send (zfs send -w), so that encrypted datasets are always sent encrypted on the wire by default? And the SSH encryption (if turned on) would basically send doubly encrypted data? What does the other "Encryption" checkbox on the right side (under the destination) do? From the video it looks like the destination dataset automatically was set as encrypted because the source was, even though, this "Encryption" checkbox was not checked in the setup of the replication. Is this checkbox only needed to turn unencrypted source datasets into encrypted ones on the target? One technical question: Does the encrypted dataset need to be locked in the moment when "zfs send -w" (or the replication task) is run in order to be sent encrypted or will it be sent encrypted even when the dataset on the source is currently unlocked (like in the example here)?
The locked / unlocked status of a volume, indeed even if it's mounted or not, has no impact on the ability to send. Keep in mind, you're sending a snapshot anyway, not the truly live data. In fact the destination server doesn't even need the encryption key for this to work, it's not important that the destination server be able to actually decrypt the volume it's receiving. ZFS send operates at the volume level, not the file level.
That is what I thought, I'm trying to use 2 of the same hardware I have for an Failover just incase trying to be like a normal SAN without like XOSAN or vSAN@@LAWRENCESYSTEMS
Is replication an extension provided by Truenas, or a built in feature of zfs? In other words, can one replicate zfs snapshots between systems that aren’t Truenas?
@@LAWRENCESYSTEMSaha thats what I am doing wrong. I really cannot touch anything in the replica? I was thinking they could both be doing snapshots and sync...
Having the two different TrueNAS systems use a different color theme makes a huge difference, thanks!
The best contents about truenas, just here, thank u. I have learn a lot since I find ur videos and forums. Thanks again
That's basically how I have my 2 truenas boxes setup. Been running them for years now. It works perfectly 😊 That said I still use core. I'll go to scale when core goes EOL
Loving these TrueNAS Scale videos! Thanks Tom. Great job.
This was what I had planned for next week. Perfect timing. Thanks... :)
I am building a second system next week to do this exact thing.
For encryption, the fact you must type a directory name to make a new directory instead of making a dataset and choosing it, is extremely confusing and impossible without this video. Thank you so much.
for more info, no matter the property you set for the designation it will not work, unless you double encrypt (encrypted designation inherited folder, and re-encrypt the snapshots sent to put in there).
doing pulls of data from a production TrueNAS to a backup one is a pretty awesome way to create hidden backups.
Just a note, I use Tailscale in subnet router mode on each side. I had to select Netcat Active Site as REMOTE for the replication task to commence.
Good tip thanks
so niceeee. Just what I needed now as I bumbed into problems. Got a secondhand HP DL380 Gen9 to get rid of my very loud and very old IBM 360. Hope I can manage the replication with this tutorial. also boot disk is failing on the old system and dont know if I can replace that so I move to the new system.
Very helpful, thank you.
Thank you for this great and detailed explanations.
Great video. I'm just about to build a 2nd TrueNAS system to do just his. One question: I plan on replicating from the current system that is still running FreeNAS-11.3-U5 to a new system running the latest version of TrueNAS SCALE. Do you think this will be a problem? What can I do instead? Thanks.
Thank you so much for this. It helped me realize what I need to do for backup.
Question I have , does the drive capacity of the backup server need to be as large as the data set that is snapshot? I ask because say all the data on my main pool is gone. 4 drives dead, z3, so the data is essentially gone. Is it possible to use these snapshots to layer back onto a new pool of drives say larger in capacity than before? That are currently blank? Or would those existing drives also need to be used?
Would it be possible to recreate from scratch like that? I’m scared to death to lose my zpool disks and need someway to replicate the data if kind of a total failure happens
Very useful video, THX.
@LAWRENCESYSTEMS A question that I see myself and others struggle with about snapshots and restoring as a backup solution is:
If I just start creating snapshots on an already established dataset that is 4TB in size, and a year from now I need to rebuild that dataset from a catastrophic failure, how would it be rebuilt from the most recent snapshot if it only reflects the last changes of say, a single pdf that was added? Wouldn't restoring with this snapshot only restore what it had detected as what changed since the previous snapshot and nothing before that?
I have a hard time wrapping my head around how a 10kb snapshot can restore a 4TB+ dataset with this.
This definitely seems to be the Ticket for Full speed no overhead Truenas-Scale backups/Replications for DATA.
I also like you can turn the backup/Replication Truenas Box into a Big JBOD Mixed Capacity Disk so I was able to put together a 35TB backup repository easily with some Misc Large/Med TB Drives i had to just serve as a backup form my ZFS Raided data. Even Still it replicated at 2.5GB/s on HDDS (Not SSDS) maxing out the 2.5GB network in my home.
I tried to do this with Veeam B&R CE via - W10/11 Backing up Proxmox VM and it was stuck Sub 1GBPS for what ever reason Didn't mater if i run the Veeam B&R CE on another HEDT or a worker on the destination off the Backup Box or with a Root install of windows just slow never broke 1GBPS I ended up giving up i just have way to much TB for it to not get a full 2.5.
Thank you for the great tutorial. I successfully replicated my dataset, but I'm having trouble replicating "ix-application." I'm running Immich (App) on my primary TrueNAS Scale server. How can I replicate it to my offline TrueNAS server?
I would like to replicate from Core to Scale. Everything is setup propperly with SSH. Problem is after just a few moments I am getting this error and I don't know which permission are wrong and how to fix that. :(
Replication "Datasetname" failed: cannot receive new filesystem stream: permission denied warning: cannot send "Snapshotname" signal received.
I've seen a lot of videos like this in setting up replication, but I haven't found anything on restoring TrueNAS from a replication after a complete failure. If system A replicates to system B, how do I restore system A from B after system A fails?
You simply reverse the process
@@LAWRENCESYSTEMSawesome. Thanks for the reply.
You do a one time, non scheduled, replication from the backup server to the main server and proceed as normal.
Okay, I had a misunderstanding of replication. When a task is completed the files are written to the other pools and appear in a filesystem the same way they are on the source end so you can interact with them. I thought replication stored it in a way where the data wasn't accessible unless recovered. I'll be using this for sure!
I have two TrueNAS scale systems one at my house, one at my parents, both running Plex. Would like to back up there to their Plex library so they can also view the movies I add and it seems like this would be a good solution. Does that sound about right?
brilliant, thank you for this well timed video. i don't suppose you have a "cloud sync" tutorial in the works?
I have some older videos on CloudBackup I might do some newer ones but the cloud backups are easier than TrueNAS Replication ones.
what do you think about FreeNas 11.2 replication to TrueNas, it's Possible ?
How can you us the Zpool iostat and other commands in Truenas Scale?
Great video!
Thanks for the video. Replication from scratch is still a bit unclear to me. Let's say the backup server has replication data from multiple TrueNAS servers. Will Replication from scratch delete all snapshots regardless of destination paths, or just the snapshots for a set destination path?
From scratch is just for that dataset path on the destination
If you manually snapshot the source and the destination, can it determine differences and copy just those or will the destination allways be wiped out? this was not clear.
You can have it send over all snapshots from source both manual and automatic, but don't make changes on the destination because they will either be over written or cause the process to fail.
Did I miss it but did you ever state that that netdata is now the built-in system reporting app?
You can load Netdata as an app
Is Raidz necessary with a setup like this if availability is not important but data integrity is?
You need at least 2 drives to validate & fix data integrity issues.
Perfect, just what I need!
So, the replication task always does a RAW send (zfs send -w), so that encrypted datasets are always sent encrypted on the wire by default? And the SSH encryption (if turned on) would basically send doubly encrypted data? What does the other "Encryption" checkbox on the right side (under the destination) do? From the video it looks like the destination dataset automatically was set as encrypted because the source was, even though, this "Encryption" checkbox was not checked in the setup of the replication. Is this checkbox only needed to turn unencrypted source datasets into encrypted ones on the target?
One technical question: Does the encrypted dataset need to be locked in the moment when "zfs send -w" (or the replication task) is run in order to be sent encrypted or will it be sent encrypted even when the dataset on the source is currently unlocked (like in the example here)?
The locked / unlocked status of a volume, indeed even if it's mounted or not, has no impact on the ability to send. Keep in mind, you're sending a snapshot anyway, not the truly live data. In fact the destination server doesn't even need the encryption key for this to work, it's not important that the destination server be able to actually decrypt the volume it's receiving. ZFS send operates at the volume level, not the file level.
@@entelin oh yeah right, the replication sends the snapshot! Then it makes perfect sense that it would always be sent encrypted
Would this be a good idea for say something like a active failover if your running your own hardware? Then on XCPNG just manual flip the connection?
There is not a simple process for that, so not really.
That is what I thought, I'm trying to use 2 of the same hardware I have for an Failover just incase trying to be like a normal SAN without like XOSAN or vSAN@@LAWRENCESYSTEMS
Is replication an extension provided by Truenas, or a built in feature of zfs? In other words, can one replicate zfs snapshots between systems that aren’t Truenas?
This is a function of ZFS but for simplicity of this video I kept it just to showing how to use it in TrueNAS.
Thanks for the reply.
sir explain from other system to nas..........in local network file sync............
Does this still work well if you use the truenas box for esxi datastores
Yes
Why is ZFS replication on target system read only by default?
So you don't mess with the backup data because if you do it will need to resend the snapshots again from the source
@@LAWRENCESYSTEMSaha thats what I am doing wrong. I really cannot touch anything in the replica? I was thinking they could both be doing snapshots and sync...
@@HermieDaddy That is why in the beginning of the video I explained the concepts of how it works and that it IS NOT a syncing system.
👍🏻👍🏻👍🏻👍🏻👍🏻
Encrypt everything... the actual cost of time/bandwidth is not that significant anymore IMHO
Smith Helen Perez Ruth Anderson Cynthia
first