Deleted files, are they gone permanently ?

[Mat1926]

I am aware of a recycle bin plugin, I don't know if I am going to add that honestly, but I am testing the system now and did delete some files from unraid mounted share in my windows 10 machine, are the files gone permanently or am I suppose to do one more step?

 

I also, transferred a 8 GB file to my unraid share, the speed is ~112 MB/sec, dual parity unraid. This is comparable to my old Raid 5 NAS appliance (speed ~114 MB/sec). I am only using 1 disk for data and 2 disks for dual parity, is this the reason I am getting high speeds similar to Raid 5 on the same network? My 2 Parity drives and my data drive are all WD Red 10 TB.

 

Adding more drives would slow down the transfer speeds correct? I was told that allowing all the drives spinning will improve the speeds in case slow transfers, where exactly is this setting and what do they call it, just in case I thought I might need it...

 

Thnx

 

 

[Squid]

17 minutes ago, Mat1926 said:

are the files gone permanently or am I suppose to do one more step?

If you didn't install the plugin, then they're gone.  If you installed the plugin, then you have to empty the recycle bin in the plugin's settings

18 minutes ago, Mat1926 said:

I also, transferred a 8 GB file to my unraid share, the speed is ~112 MB/sec, dual parity unraid. This is comparable to my old Raid 5 NAS appliance (speed ~114 MB/sec). I am only using 1 disk for data and 2 disks for dual parity, is this the reason I am getting high speeds similar to Raid 5 on the same network? My 2 Parity drives and my data drive are all WD Red 10 TB.

With 1 data drive, unRaid automatically recognizes this and automatically puts the entire system into "Turbo Mode (Reconstruct Mode)" (At least it does with single parity.  With dual parity it still should since I presume the 2nd parity drive at that point is still a mirror of the first with only 1 data drive)

 

With more than 1 data drive, writes will slow down (probably to ~60-80MB/s depending upon a number of factors) as the system won't be in turbo mode by default and will suffer the penalties imposed by the read / modify of the parity data.  You'll get back up to 100+ (limited by network speed) by enabling Reconstruct Mode in Settings - Disk Settings.

[JorgeB]

6 hours ago, Squid said:

With 1 data drive, unRaid automatically recognizes this and automatically puts the entire system into "Turbo Mode (Reconstruct Mode)" (At least it does with single parity.  With dual parity it still should since I presume the 2nd parity drive at that point is still a mirror of the first with only 1 data drive)

With a single data drive unRAID works like raid1, or dual mirror with dual parity, so it's much faster than normal parity, though with turbo write enable when using more data drives speeds can be similar.

[Mat1926]

6 hours ago, Squid said:

If you didn't install the plugin, then they're gone.  If you installed the plugin, then you have to empty the recycle bin in the plugin's settings

With 1 data drive, unRaid automatically recognizes this and automatically puts the entire system into "Turbo Mode (Reconstruct Mode)" (At least it does with single parity.  With dual parity it still should since I presume the 2nd parity drive at that point is still a mirror of the first with only 1 data drive)

 

With more than 1 data drive, writes will slow down (probably to ~60-80MB/s depending upon a number of factors) as the system won't be in turbo mode by default and will suffer the penalties imposed by the read / modify of the parity data.  You'll get back up to 100+ (limited by network speed) by enabling Reconstruct Mode in Settings - Disk Settings.

 

2 minutes ago, johnnie.black said:

With a single data drive unRAID works like raid1, or dual mirror with dual parity, so it's much faster than normal parity, though with turbo write enable when using more data drives speeds can be similar.

 

Great, I found the reconstruct write in the Tunable (md_write_method). I was told the only downside is that all the drives will be spinning simultaneously which means more power/heat, and maybe unconfirmed reports of degradation of the mechanical hard drives on the long run. Your thoughts about this?

 

If I wanted the HDDs to spin down if not used for sometime, where is that option? Is this any good at all?

 

Can I mount my WD PR4100 32 TB NAS appliance (Raid 5) in unraid and transfer the files locally instead of over the network? I have 2 of these, one with 23 TB of data, and the other with 9 TB of data...

 

If the only option is to transfer the files over network, do I need to md5 check each file after the transfer, or that is not necessary? because when I'm done transferring the files from each appliance I will use the 8 TB Red HDDs in my unraid, so I would like to make sure that I do have an exact copy of my data. I use Total Commander to copy/move files, and it got a feature to do md5 tests upon moving files...

 

Thnx

[JorgeB]

4 hours ago, Mat1926 said:

I was told the only downside is that all the drives will be spinning simultaneously which means more power/heat, and maybe unconfirmed reports of degradation of the mechanical hard drives on the long run. Your thoughts about this?

Power and heat yes, better for reliability it's debatable, most believe, including myself, that a disk running 24/7 will last longer than a disk with lots of spin up/down cycles.

 

4 hours ago, Mat1926 said:

If I wanted the HDDs to spin down if not used for sometime, where is that option? Is this any good at all?

Settings -> Disk Settings

 

Disks will still spin down with turbo write enable when idle, and only the needed disks will spin up for reads.

 

4 hours ago, Mat1926 said:

Can I mount my WD PR4100 32 TB NAS appliance (Raid 5) in unraid and transfer the files locally instead of over the network? I have 2 of these, one with 23 TB of data, and the other with 9 TB of data...

You can mount them with the UD plugin but the transfer will always be over LAN.

 

4 hours ago, Mat1926 said:

If the only option is to transfer the files over network, do I need to md5 check each file after the transfer, or that is not necessary?

It's not mandatory as ethernet has crc checking on all transfers, still if you want to be extra safe you can do it, same applies to local transfers.

[Mat1926]

2 hours ago, johnnie.black said:

Power and heat yes, better for reliability it's debatable, most believe, including myself, that a disk running 24/7 will last longer than a disk with lots of spin up/down cycles.

 

Settings -> Disk Settings

 

Disks will still spin down with turbo write enable when idle, and only the needed disks will spin up for reads.

 

You can mount them with the UD plugin but the transfer will always be over LAN.

 

It's not mandatory as ethernet has crc checking on all transfers, still if you want to be extra safe you can do it, same applies to local transfers.

 

Thanks for all the help...Would you please look at these screens and confirm that my cards are properly setup...

 

 

 

*EDIT*

I have a m.2 pcie ssd unrecognized device. What I want is to use it for virtualization/docker support and perhaps move some of the contents that are not needed like plugins for instance from my main array. Is it possible to use it for that purpose only, or it will also be used to cache any transfers?! I prefer to copy/move the files directly to the array w/o any caching...

 

Thnx

[Mat1926]

One more question. I started copying ~9 TB of data to ~9.3 TB unraid free space. So when I am done I will have like 300 GB only of free space! Then after making sure that all data exists and okay, I will use the old 4 x 8 TB HDDs in my unraid system. When I do that, would the system transfer some of the files from the almost full HDD to the new ones? My setup is 2 Parity and 1 Data at the moment, all 10 TBs each...

[itimpi]

17 minutes ago, Mat1926 said:

One more question. I started copying ~9 TB of data to ~9.3 TB unraid free space. So when I am done I will have like 300 GB only of free space! Then after making sure that all data exists and okay, I will use the old 4 x 8 TB HDDs in my unraid system. When I do that, would the system transfer some of the files from the almost full HDD to the new ones? My setup is 2 Parity and 1 Data at the moment, all 10 TBs each...

UnRAID never does any automated transfers of files to new disks - it always needs to be a manual process initiated by the user.  There is a plugin called unBalance that can help with doing this.

[Mat1926]

26 minutes ago, itimpi said:

UnRAID never does any automated transfers of files to new disks - it always needs to be a manual process initiated by the user.  There is a plugin called unBalance that can help with doing this.

 

Is it necessary to balance the files, or it is totally fine just to use the system as is...one full disk and 4 free ones?

[JonathanM]

6 minutes ago, Mat1926 said:

 

Is it necessary to balance the files, or it is totally fine just to use the system as is...one full disk and 4 free ones?

Totally fine, some (including me) would say preferable. Writes to an empty drive are faster, so why would you use up the fastest space on each drive by filling it partway? As long as the share is configured correctly, new writes will go to whatever drive you want.

 

The only exception would be if there are going to be lots of files added that need to physically stay on the same drive for spin up minimization, then you need to keep enough free space to account for the added files.

[JorgeB]

7 hours ago, Mat1926 said:

 

Thanks for all the help...Would you please look at these screens and confirm that my cards are properly setup...

They look fine.

 

7 hours ago, Mat1926 said:

*EDIT*

I have a m.2 pcie ssd unrecognized device. What I want is to use it for virtualization/docker support and perhaps move some of the contents that are not needed like plugins for instance from my main array. Is it possible to use it for that purpose only, or it will also be used to cache any transfers?! I prefer to copy/move the files directly to the array w/o any caching...

Yes, you can use with the Unassigned Devices plugin

[Mat1926]

15 minutes ago, jonathanm said:

Totally fine, some (including me) would say preferable. Writes to an empty drive are faster, so why would you use up the fastest space on each drive by filling it partway? As long as the share is configured correctly, new writes will go to whatever drive you want.

 

The only exception would be if there are going to be lots of files added that need to physically stay on the same drive for spin up minimization, then you need to keep enough free space to account for the added files.

 

Thnx a lot for this...I appreciate it.

[Mat1926]

5 minutes ago, johnnie.black said:

They look fine.

 

Yes, you can use with the Unassigned Devices plugin

 

The system will not copy the files first to the cache drive? No need to disable anything or adjust the settings?

[JorgeB]

8 minutes ago, Mat1926 said:

 

The system will not copy the files first to the cache drive? No need to disable anything or adjust the settings?

No unassigned devices are shared outside the array, they use a different path, /mnt/disks/your_path

[Mat1926]

1 minute ago, johnnie.black said:

No unassigned devices are shared outside the array, they use a different path, /mnt/disks/your_path

 Strange, I was watching a tutorial on youtube, and the guy said that one of the reasons of having cache drives is to transfer the files first to the cache drives, then the server when it is not busy will transfer the files to the array. So I thought I have to disable that feature and use the cache disk for my plugins etc only....

[JorgeB]

2 minutes ago, Mat1926 said:

 Strange, I was watching a tutorial on youtube, and the guy said that one of the reasons of having cache drives is to transfer the files first to the cache drives, then the server when it is not busy will transfer the files to the array. So I thought I have to disable that feature and use the cache disk for my plugins etc only....

I understood you already had a cache device and that was an extra device, if you don't have cache just assign it as cache, you can then select the appropriate type of cache use for each share, see the WebGUI help or the FAQ for the description of the 4 modes available.

[Mat1926]

1 minute ago, johnnie.black said:

I understood you already had a cache device and that was an extra device, if you don't have cache just assign it as cache, you can then select the appropriate type of cache use for each share, see the WebGUI help or the FAQ for the description of the 4 modes available.

 

Okay thnx, will do that....