Where snapshots are stored?

[OmarNoah]

Good morning to all.

I'm having trouble with my freight service, and even though I search the net for some information, I can't get hold of the location of the snapshots.

In short, my problem is the following.

The snapshots consumed the disk space that I had available, and it is generating serious problems for me at the moment.

At the moment, I have only 3.5 gb available, I tried to perform the information reprocessing of some of the datasets, but despite this, there is no evidence that the available space increases. But when I delete snapshot logs, the available space immediately increases. I have tried several things, add a 250 GB disk, to try to add it to the pool, but I already have 5 1TB disks, and it does not allow me to add a 250 GB disk to that pool (I found in the documentation that it did not allow me .) I tried to clone the snapshots that occupied the most space, in a new pool with the 250 GB disk, but it does not allow me to clone them if they are not in the same volume.

The documentation is clear in saying that you can only add disks of the same size if you want to add space to that pool.

My idea would be to expand the unit where the snapshots are saved, or to save the snapshots on the additional disk I had, but I have not found how to carry out that process, how to tell the system to save the snapshots on a certain disk. I would appreciate very much if you could help me.

 

[garm]

S.T.O.P!!! (Sit, Think, Observe, and Plan). Your data is in serious danger atm...

You are risking your server due to gross noobism. Snapshots aren’t stored anywhere other then where your data is.

3 GB on a 5 TB pool is 99.9..% utilization.. you will see performance degradation after 90%, and a warning after 80%.

snapshot don’t consume data per say, they prevent modified or deleted data from being wiped off the disk, so that it’s still retrievable.

also, you don’t seam to have any redundancy in this pool..

Your best bet now is to offload the data to a properly configured pool... I fear you won’t be able to import this pool if you reboot the machine.. it might work to import it read only, but I would not risk it. Build a new machine and a proper pool, then transfer the data over and rebuild this pool properly.

 

[danb35]

Now, if you have lots of snapshots you want to delete, a script can help there. Something like this:

Code:

#!/bin/bash
for std in `zfs list -H -o name -t snapshot | grep -- 'auto'`
do
    echo "zfs destroy $std"
    zfs destroy $std
done

...will delete every snapshot with "auto" in the name (which should cover any snapshots created by automatic tasks on your system).

 

[HoneyBadger]

You need to find a way to get another logical copy of that data right now; from the looks of things, your pool is set up as striped (5x 1TB drives, 4.8T total) which means that any one drive failing will cause your whole pool to become inaccessible.

Whether this is a USB drive on a Windows machine or something else, you need to get a second copy right away.

Once that's done, start working towards the snapshot deletion to free up space - but your pool needs to be re-architected in a redundant manner.

 

[OmarNoah]

[QUOTE = "garm, publicación: 609294, miembro: 77744"]
¡¡¡DETENER !!! (Siéntese, piense, observe y planifique). Tus datos están en grave peligro cajero ...

Estás arriesgando tu servidor debido al gran noobismo. Las instantáneas no se almacenan en ningún otro lugar que no sea su información.

3 GB en un grupo de 5 TB equivalente al 99,9% de utilización ... verá una degradación del rendimiento después del 90% y una advertencia después del 80%.

Las instantáneas no consumen datos, por ejemplo, evitan que los datos modificados o eliminados se borren del disco, por lo que aún se pueden recuperar.

Además, no parece tener redundancia en este grupo.

Su mejor opción ahora es descargar los datos a un grupo establecido correctamente ... Me temo que no podrá importar este grupo si reinicia la máquina ... podría funcionar importarlo de solo lectura, pero no me arriesgaría eso. Construya una nueva máquina y un grupo adecuado, luego transfiera los datos y reconstruya este grupo correctamente.
[/ CITAR]


Hola y gracias a todos por responder. Tengo el servidor replicado y creo que todo está bien por ese lado, lo que no quiero es que se generen los problemas que me dices. ¿Podría alguien decirme cómo puedo validar que los discos de este servidor tengan redundancia?

 

[danb35]

Yo siento, pero no se habla Español.

 

[garm]

Nä inte jag heller

 

[ornias]

Spaans hebben we hier dus echt helemaal niks aan.

 

[danb35]

Missed opportunity--I should have written:

Das kommt mir Spanisch vor.

 

[OmarNoah]

Hello everyone, the solution I found was to add the 250GB disk to a new pool, and pass information there so that the server can continue to operate normally for a little longer, while debugging is carried out. I offer an apology for having written in Spanish, I did not expect so much show, but hahaha, thanks anyway. I learned a lot from your observations. Happy rest of the day.

 

[danb35]

Be aware that this is still a very temporary solution--as Honey Badger (who doesn't give a sh*t) says, you have no redundancy with your pool, which means that if (when) a disk fails, you will lose all the data stored there--and given that these are 1 TB disks, I'd speculate they're fairly old already, and thus failure would be expected sooner rather than later.

 

[OmarNoah]

Hi danb35, ommm, I understand much better what you are telling me, the recommendation would then be to transfer all the data to a removable hard disk for example, and remount the server with RAID 1 or RAID 5.

(I have a question, how can I validate what type of RAID I have at the moment.)

That is what I think I understand.

As I mentioned earlier, the server is currently replicating to another server.

But with the above, a question arises, that server is a virtualized server with an external provider.

Is it recommended that it be an independent physical server?

 

[danb35]

how can I validate what type of RAID I have at the moment.

You can see that on the pool status page in the GUI. If you let us know which version of FreeNAS you're using, we could tell you how to get there, or you can check the manual (use the Documentation link at the top of this page).

Is it recommended that it be an independent physical server?

In most cases, yes. FreeNAS (and ZFS generally) expects to have direct access to the hard disks so that it can monitor their condition. It's possible to arrange for this in a virtual environment, but somewhat tricky.

 

[OmarNoah]

You can see that on the pool status page in the GUI. If you let us know which version of FreeNAS you're using, we could tell you how to get there, or you can check the manual (use the Documentation link at the top of this page).

In most cases, yes. FreeNAS (and ZFS generally) expects to have direct access to the hard disks so that it can monitor their condition. It's possible to arrange for this in a virtual environment, but somewhat tricky.

My versión is FreeNAS-11.2-U3.

 

[HoneyBadger]

My versión is FreeNAS-11.2-U3.

Storage - Pools - (choose the Gear icon in the top right) Status

Post the screenshot there, but from your description and the total size you have a "RAID0" equivalent - all of your data is spread across all drives.

 

[ornias]

Well, the good thing with his setup... If he needs space he can just rip out one drive and he'll have all the free-space he would ever need