ReadyNAS 312 swapping HDD for SSD, filesystems, data transfer

After you remove the old HDD1 (which is currently your primary volume -- which sdx device it is has no bearing, and can even change between boots), you will have no primary volume and the old one will still show up as missing.  You have two options, depending on what you intend to do with it, to force the NAS to make the other HDD primary:

Before you remove it, you can EXPORT it.  If you do that, it cannot be used as a primary drive -- it has to be imported by another one (the one on HDD2 or any other).

After you remove it, you can DESTROY the "phantom" volume.  That will cause the NAS to forget it.  But since it's already removed, it'll do nothing to the actual volume so it could still be used as a stand-alone primary (but it will remember the HDD2 volume, so you'll have to deal with that).

Well, you really have a third option of destroying before removal, but that's contrary to what I think you want to accomplish.

Doing both of the first ones may be the best path forward.  EXPORT HDD1, thus making HDD2 primary, then re-import HDD1 (by booting with the two installed).  HDD1 will then have whatever the NAS does to show it was "EXPORTED" (I've not figured that out) un-done and be available to move things to HDD2, but it won't take over as primary.  After you've done that, you can then remove it and DESTROY the phantom and now have both HDDs in a configuration that they can be a stand-alone volume. 

It is unusual to see sdb in the list.

---

ddewrbbdg wrote:

 

Regarding root partition i have found few topics, that show and dont recommend messing with that or resizing it.

I will try to find and truncate or delete logs right after and find big files consuming a lot of space. Also some apps could be it.

 

---

Part of the problem is here: 

1.4G /etc

Normally it is around 12 MB

The best way to explore the OS partition is to remount it as /mnt, and then explore /mnt.  That trick prevents the various mount points from obscuring the actual space use.

mount --bind / /mnt

You can undo the mount with the umount command when you are finished.

Sandshark_Sensei

Thank you, for the guide. I have read it twice and i will have to do it more. Difficult to understand all the steps. After i transfer remaining data, i wanna swap slot1 - old - HDD1 for empty SSD that already has system (OS), JBOD and Flex Raid, cause it was connected already before in slot 2, and i want leave slot2 - new - HDD2 untouched further.

Would be also great to keep disconnected/removed HDD1 with data, as is now, just for case.. anything goes wrong as backup drive with 4TB of data. I can format / destroy it anytime later from any PC.

vs.

Or should i do this Destroy - before i remove disk? Does it take away only root partition or data partition too, so both?

Regarding SSD that has been connected before as secondary, should i connect it as it is, or format it using windows before.???

------------------------

Regarding folders / drive names.. i have chosen name "sdb" for new HDD2, cause i was reading some guides regarding naming drives in linux 😞 and intented to call SSD "sda" as they go in the following order: slot1 SSD, slot2, HDD2. I dont know if im gonna be able reach this.

------------------------

LOOKS LIKE BACKUP TASK is going to be completed finally after 10 hours. So gonna do next steps then. I think scheduled backup turning off / removal has helped.

REGARDING consumed space in ETC.. have to do further examination and research. Listed directories, with sizes, nothing unusual, listed files same, listed files recursively, have to put it into text file, some good text editor and browse it line by line again. I think one of commands was showing only .. in /etc/ folder "." is consuming 1,4GB.. but what it could be? There were some cache files and im still copying. Maybe that could be source of problem. Gonna see

Sandshark 

I have read what u recommend. My question is...

I got now primary in slot 1 = HDD1, that i want remove once for all and scrapp / dispose.

HDD2 the new disk with data .. finally - in slot 2.

I want disconnect HDD1 from slot 1 and replace it with SSD (that has been already connected to slot 2 few times and is formatted as JBOD with this very same system). And then make this SSD primary. I dont wanna make primary the new HDD2.

Should i make the export - import part? Like is it possible to export, remove slot 1 HDD, then insert SSD, then import to SSD? Or would this crash. As what i have seen, export says it destroys data too, so it wouldnt be possible to transfer it from HDD2 to SSD in slot 1 later, if id put it on HDD2.

Like will NAS still work with removed old primary HDD only with HDD2 meanwhile until i connect SSD into slot 1? Then somehow making this SSD primary? I dont understand the process of export import much. Thank you

---

ddewrbbdg wrote:

 

I want disconnect HDD1 from slot 1 and replace it with SSD (that has been already connected to slot 2 few times and is formatted as JBOD with this very same system). And then make this SSD primary. I dont wanna make primary the new HDD2.

 

---

Once you destroy or export HDD1, HDD2 becomes primary.  All that means in practice is that .apps on hdd2's data volume is mounted as /apps.  If the SSD is in slot 1, the NAS will boot from the SSD even if HDD2 is primary.

Sandshark:  After SSD is synced, and .apps is copied onto the SDD2 volume I am wondering if this would work???

1. export HDD2 - which will make the SSD primary
2. remove HDD2
3. destroy the data volume for HDD2
4. power down 
5. reinsert HDD2
6. reboot (which I think should import HDD2's data volume back, but SSD should remain primary)

Any thoughts on this?  ( ddewrbbdg:  The risk here is that the import in step 6 might not happen.  Sandshark has played with export more than I have, which is why I'm tagging him on this approach). 

---

ddewrbbdg wrote:

As what i have seen, export says it destroys data too,

---

Export doesn't destroy data - where are you seeing this?

---

ddewrbbdg wrote:

[post 16] Also found /dev/md .. md0 ,, md1 .. md125 .. md126 .. md127 but cant enter as directory using cd or dont know how to work with that.

 

---

Just noticed that neither Sandshark nor I answered this comment.

RAID (mdadm) creates a virtual disk from for each RAID group it creates.  

• md0 is the virtual disk for the OS
• md1 is the virtual disk for swap
• md125, ... are virtual disks for the various data volumes on your hard drives.  Even with JBOD, the NAS is creating a virtual disk with mdadm for each volume.

With Linux you can't use cd to access the raw disk (no matter whether the disk is physical or virtual).  You have to create a file system on that disk (btrfs in your case), and then mount it.  

• md0 is mounted as the root / 
• md1 is not mounted, as it does not have a file system installed on it
• Each volume has a placeholder folder on the root, which has the same name as that volume. md125, etc are mounted to that folder.

With HDD2 as primary and the SSD as secondary, you can simply EXPORT HDD2, thus making the SDD primary, then re-import HDD2.

Destroying a volume does not destroy the OS partition.  It does make some changes so it no longer tries to mount the destroyed partition.  I've not investigated the extent of those changes.

Note that there is always some risk whenever you do this, so make sure you keep your backup up to date.  But I did a lot of experimenting in this area -- several in a row with the same drives/volumes (albeit with an earlier version of the OS, probably 6.9.5) -- and only had a problem because the "sandbox" drives I was using were ones retired from a NAS due to re-allocated sectors and one failed.  But, even new drives can fail.

---

Sandshark wrote:

With HDD2 as primary and the SSD as secondary, you can simply EXPORT HDD2, thus making the SDD primary, then re-import HDD2.

Can you refresh me on exactly what the re-import process would be?  It's not something I've ever needed to do on my own NAS.

Is it just rebooting?  Or more involved?

Import is automatic if you boot with the previously exported volume installed together with a primary volume.  That's why you won't find a process for it anywhere.  Note that the volume name and all share names cannot be the same as on an existing volume.  As far as I can tell, it simply doesn't show the duplicate-named shares in the GUI or via any sharing protocol, so you can access them via SSH if needed.  But the best path is to re-name any on the primary that will be duplicated, then import, and then fix whatever naming you need to fix.  The only solution to having the same volume name is the very complex volume re-naming process I outlined here: httChanging-the-volume-name-on-an-OS6-based-NAS. 

You didn't ask, but you can import with either volume being degraded (e.g. just one drive of a RAID1).

Local language translation says this / i translate back into english:

Type EXPORT to confirm and explanation below.

 >>> "By destroying this volume will be all the data erased permanently on this volume." <<<

It's unclear from above to me, whether this takes all partitions down, or just the OS partition.

--------------------------------------------------------------------------------------------------------------------------

I'm still thinking how to make empty SSD with mirrored OS partition set as JBOD primary easier or directly.

What if i remove HDD2 now... connect SSD to slot 2 instead, then export old HDD1, import to SSD.

Then swap SSD from slot 2 to slot 1 and finally connect HDD2 back to slot 2?

At worst case i can afford lose OS partition, or reinstall OS ... the best case = in SSD. But what i cant afford is losing one of hard drives in JBOD regime with the data. I am willing use HDD2. No more life for HDD1.

// Another thing im thinking of is getting SATA to USB 3.0 reduction including power. For connecting other or spare drive(s) for backup or if anything fails. Question is .. just quick idea. Could NAS boot from this device via USB3.0?

---

ddewrbbdg wrote:

I'm still thinking how to make empty SSD with mirrored OS partition set as JBOD primary easier or directly.

 

---

Could be a language barrier here, but we have already told you that.

1. Create a data volume on the SSD, and copy .apps to it.
2. Export HDD2
3. Reboot

EXPORT is not the same as "Destroy", so I am puzzled by your warning message.

Ok.. question is.. i dont have SSD attached now only old HDD1 with HDD2. I cant just remove boot HDD1 and hotswap SSD there? 🙂 bad idea right?.. for copying .apps to ssd and reboot. Looks logical to me remove HDD2, insert SSD copy files there.. then SSD to 1st slot, = with HDD2 to second slot, exporting HDD2 then. Right? I got only 2 chassis device ReadyNAS 312 and no SATA-USB reduction.

Yeah it could have been language barrier. Also ReadyNAS is translated somehow by someone. Maybe the translator was also human and didnt understood export process well too, or tested the export process and it took down his data, before he learned more info.

Im just asking multiple times, cause the data. And on previous page in post 7 and others i was adviced to export HDD1 that doesnt seem logical to me. Exporting HDD2 is.

---

ddewrbbdg wrote:

Ok.. question is.. i dont have SSD attached now only old HDD1 with HDD2.

First

1. Make sure the .apps folder on HDD1 has been copied to .apps on HDD2 (don't delete .apps on HDD1)
2. Export HDD1
3. Remove HDD1 and set it aside for now
4. Destroy HDD1 if it still shows up on the volume page.

At this point, slot 1 should be empty and HDD2 should be in slot 2.  The system should boot normally, and all your apps should continue to work.  I recommend testing this before proceeding.

The data and OS also remain on HDD1 (since you destroyed the volume after the disk was removed).  So it serves as a backup - meaning you can power down the NAS, and put HDD1 back in slot 1 by itself.  The NAS will boot as it did before.  It is safe to test this, just

1. power down
2. put HDD1 in slot 1
3. remove HDD2
4. boot

Make sure the NAS is powered down when you remove HDD1 and put HDD2 back into slot 2.

After you've checked all this then 

1. Boot the NAS with slot 1 empty and HDD2 in slot 2
2. Insert the SSD into slot 1 and create a volume on it.  Given the history this might require formatted the SSD in the NAS.
3. copy .apps from HDD2 to the SSD volume.  (No need to delete .apps on HDD2, and I suggest you leave it there for now)
4. Export HDD2
5. Reboot 

Now the SSD should be primary.  You can check that with ssh. 

1. First identify the virtual disk that was created for the SSD volume with cat /proc/mdstat.  This would be either md126 or md127.  Since the SSD is smaller than HDD2, you can easily tell which is which.
2. Second use mount | grep apps to show you which virtual disk is mounted as /apps.

Note you can also check this by looking at mdstat.log and mounts.log in the log zip file.

No more reading no more posting. I m gonna try it. Now u gave me guide how to transfer OS from HDD1 to (via) HDD2 to SSD. I think that, from HDD1 to SSD in slot 2 will be faster (requiring only 1 transfer). Then i just change drives in slots to final positions. I think and hope ReadyNAS OS is little smart too and hope. Of course im copying .apps to all drives now.

---

ddewrbbdg wrote:

 I think that, from HDD1 to SSD in slot 2 will be faster (requiring only 1 transfer). 

---

I don't understand.  If HDD1 is in slot 1 now, then the simplest thing is to export HDD1 and then add SSD to slot 1.

If you are concerned about copying .apps twice, then you are overthinking this.  It shouldn't take long to copy .apps

What you want to avoid is data loss, and having to reinstall the apps from scratch.

Long story short..

Left old HHD1 + connected SSD, copied .apps where it should be (present on all the disks and at data folder), copied pihole configs.

Then removed old HDD1. Connected SSD to slot 1, started system..

Received the following erorr message: for every installed app (5 times):

"System: App [appname] with wrong config has been found. Forbidding its configuration page".

Then i have seen apps no more. Tried apps http url configs, not available anymore nor running.

Then started with both drives SSD + HDD2.

Double checked and both drives contain .apps, where it should be, but was not present in /data/ folder. So copied there too. Restarted again and still no apps, they are gone. Some clue how to fix? All data except this are ok. Any chance to renew apps, or need to reinstall?

mount | grep apps ............... shows nothing

Another problem is i see no shares in LAN network, SAMBA. Tried turn off and on again SMB shares. Folders are there, with the files. But during SMB turning off i get this message and error code 1002030001: Commit failed.

Third .. question.. is ReadyNAS able to boot from USB3? I gues not cause need linux controllers?

---

ddewrbbdg wrote:

 

Left old HHD1 + connected SSD, copied .apps where it should be (present on all the disks and at data folder), copied pihole configs.

 

Then removed old HDD1. Connected SSD to slot 1, started system..

 

---

Is there a reason you didn't follow my steps?

Have you tried

1. powering down
2. putting HDD1 back
3. rebooting

It wasnt that much clear to me as guide from Sandshark Sensei, that appeared shorter and easier. Also i though as him that ReadyNAS will handle this. I will try with HDD1 ... in slot 1 in few hours, when i'll find some free time. May i leave HHD2 still connected? Or first launch from HDD1 without connected HDD2 or any secondary drive?

Then i will try make all the steps according your guide. I was really exhausted yesterday.. instructions look clear to me.

But what i was thinking of.. and i probably would do or need in the future (cause messing with packages and apt-install due custom apps) is.. would be possible to reinstall ReadyNAS on primary SSD from scratch somehow, just clean install, without losing data on HDD2? I know i can do factory reset, but it destroys all the data for sure?

---

ddewrbbdg wrote:

It wasnt that much clear to me as guide from Sandshark Sensei, that appeared shorter and easier.

 

---

FWIW, Sandshark never gave you a step-by-step guide.  He just sketched out what you needed to do.

I'm not sure exactly what you did - your summary doesn't have enough detail for that.  But I am thinking that looking for shortcuts got you into trouble. 

So I suggest starting over, powering down and booting with only HDD1 in slot 1.  Make sure that it boots, your data is all available, and that the apps/packages all work before going on to the next step.

---

ddewrbbdg wrote:

But what i was thinking of.. and i probably would do or need in the future (cause messing with packages and apt-install due custom apps) is.. would be possible to reinstall ReadyNAS on primary SSD from scratch somehow, just clean install, without losing data on HDD2? 

---

You can just export HDD2 and then remove it. 

After do a factory reset on SDD, switch to FlexRAID, and reinstall your apps. (That will require fixing apt, using the steps in the link I gave you above).

When that is working, power down, reinstall HDD2 and reboot.

Is there some confusion as to what "EXPORT" is?  It doesn't mean just removing the drive, it's a specific function on the Volume menu gear icon.  While a DESTROY can be done with the volume already removed, EXPORT cannot, because it makes changes on the exported drive.

There may also be some confusion as to what a "primary" volume is.  It has nothing to do with what slot something is in.  It is something ReadyNAS OS labels the drive that contains the /apps and /home shares and configures systemd to mount them.  In order to get the NAS to change the primary drive, you must EXPORT or DESTROY that primary volume.

Your /etc/systemd/system/multi-user.target.wants/apps.mount probably points to the wrong volume for /apps, and you can fix that manually, but it won't fix everything that lack of a proper primary volume will cause.  There are many other mounts and configurations that are affected.

If you take a look at Changing-the-volume-name-on-an-OS6-based-NAS, you'll see everything I found that's affected by a change in the volume name, which is effectively what you've done, but there could still be more associated with the primary volume I've not investigated. 

I have tried what Sandshark recommended, there is really something wrong with shortcuts, but the file has right content and values inside of it.

/etc/systemd/system/multi-user.target.wants/apps.mount

[Unit]
Description=Apps Directory
Before=readynasd.service
ConditionPathIsDirectory=/data/.apps

[Mount]
What=/data/.apps
Where=/apps
Type=none
Options=bind

[Install]
WantedBy=multi-user.target

Folder /data/.apps/ exists with folders with apps inside of it. Except folder /apps/ was physically empty. Using ssh via root user + then login as root via su - dont have sudo, didnt help.. copying into that directory gives permission denied no matter if i copy whole apps or folder by folder. Also i cant create empty folders inside /apps/ ... so is dead end here.

/// regarding confusion about export - local translation says it destroys all the data, i think translator has been underpaid, cause there should have been noted, it could destroy all the data, if you do any mistakes.

/// i understand what volume is something like partition but above it.

Looks like altough HDD1 putting back would probably work, i need or / want try clean OS reinstall / factory reset and clean install of OS to SSD. Cause probably gonna need this in a future too, or hopefully not.

Stephen_B: Are u 100 percent sure, that if I disconnect HDD2, leave only SSD connected, do factory reset / clean OS reinstall

and after succesful OS install to SSD, after i connect HDD2 again, it will be readable and available right after connecting with all the data from before, or does it destroy this data too (like you said there are some records of "volumes" in webadmin, that will be deleted with OS reinstall).

Of course im leaving HDD1 as backup with data now.

---

ddewrbbdg wrote:

 

Folder /data/.apps/ exists with folders with apps inside of it. Except folder /apps/ was physically empty.

The fact that /apps was empty means that /data/.apps wasn't mounted.  Did you try rebooting the NAS?

---

ddewrbbdg wrote:

 

Stephen_B: Are u 100 percent sure, that if I disconnect HDD2, leave only SSD connected, do factory reset / clean OS reinstall and after succesful OS install to SSD, after i connect HDD2 again, it will be readable and available right after connecting with all the data from before, or does it destroy this data too (like you said there are some records of "volumes" in webadmin, that will be deleted with OS reinstall).

 

---

While I didn't try it, I gave you the correct process. 

I said to export HDD2 (using the volume settings wheel) and then remove the disk. The import process (done after the reset) will restore the config files for the volume.  FWIW, that is what the export/import procedure is designed to do.  Export sets up the volume so it can be imported on another ReadyNAS (or in your case, the same ReadyNAS later on).

I have restarted once more NAS and still empty apps folder. I think there have to be messed up more services than only smb and this one.

I think we are walking in round.

If i remember you said reinsert HDD1 export to HDD2.. then remove HDD1, insert SSD instead and export HDD2?

I go according your previous guide in hour or two.

---

ddewrbbdg wrote:

 

If i remember you said reinsert HDD1 export to HDD2.. then remove HDD1, insert SSD instead and export HDD2?

 

---

What I suggest first is to go back to your starting point.  That would be done by

1. powering down the NAS
2. Putting HDD1 in slot 1
3. Leaving slot 2 empty.
4. booting up.

Then confirm that HDD1 (when booted alone) is fully working.

After that,

1. power down
2. add HDD2 to slot 2
3. reboot

If that imports properly, then copy .apps from HDD1 to HDD2 again.  If it doesn't import, then format it, create the volume, create the shares, etc again.

Then

1. export HDD1
2. power down
3. reboot with slot 1 empty. 
4. Check that /apps is properly mounted.  Use mount | grep apps from ssh

If /apps is ok, then

1. insert SSD into slot 1
2. format it
3. recreate the volume on it
4. copy .apps from HDD2 to SSD
5. export HDD2
6. reboot
7. check that /apps is properly mounted (pointing to the SSD).

If these steps don't work as expected, then I suggest stopping and posting what you are seeing here.  The checks are intended to make sure everything is ok before moving on to the next phase.

Then you can follow the steps in post 32.

Good news. What i did was no OS transfer, but clean reinstall + disconnected HDD2 and connected back after.

1] removed HDD2 from slot2

2] inserted SSD and made clean install including factory reset

3] then i have connected HDD2 and it was adopted flawlessly

- i cant find out from which drive system boots, its taking still too much time like 3 minutes or little more, i would expect better performance, altough restarts are not frequent.

First i took 6.10.10 version latest, awared that there could be problems with apps installation. idk if it was direct cause, many packages were missing, so i switched to 6.10.9 then just for being sure 6.10.8 and made recommended changes in sources.list specifying working links to repos.

I think all ReadyNAS OS all parts work, services too. There was some problem with delay when using ssh to login, during verifying username and password, it took about 1-2 minutes each before it finally logged in. I think it was some kind problem with ip address, temporarily occuring, now looks solved by swapping between dhcp and static.

Verified network address, can detect OS updates correctly, so it is ok including default gateway and dns server.

I HAVE ACHIEVED installing all 3 apps=php,mysql,phpmyadmin including keyrings from different dates and some certificates downloading. Also had to install "whiptail" for dialogs manually. During packages install, there is still some problem with "nfs-kernel-server" would like to reconfigure it, or reinstall but it doesnt work. Now i cant install some packages dues this error anymore... 😞 nor

THE MAIN PROBLEM is, pi-hole docker from rnxtras package installation does not work neither from gui webadmin, with no error spec, nor from ssh, where it give pihole-FTL not present. Usually it has installed about 25-28 packages from gui or ssh, now i cant find out what is wrong or how to continue and solve this lack of package pihole-FTL.

If you'd know solution or could help pls.. i d really appreciate it.

---

ddewrbbdg wrote:

 

- i cant find out from which drive system boots, its taking still too much time like 3 minutes or little more, i would expect better performance, altough restarts are not frequent.

---

As we suggested before, once RAID starts, the NAS will begin using both disks.

---

ddewrbbdg wrote:

 

First i took 6.10.10 version latest, awared that there could be problems with apps installation.

The only issue with 6.10.10 is that it removes the ability to install apps or deb packages from the apps page of the web ui. One work around is that you can install deb packages as if they were firmware updates, or you can of course use ssh.

If you do reinstall firmware, double-check that the fixes to the apt configuration haven't been over-written.