There are a few of these. The big ones have to do with migration because you have to remember a bunch of things. The first is that when you are moving to a new server and want to keep your old one, you have to break all your replications. So the main plan is to:
- Install your new server with new drives replicate over to it
- Then take your now backup server and replicate it back. Seems silly to do it this way, but that’s all I’ve been able to figure out.
One of my problems has been that I have completely populated 12-bay drives, so to do this replication, I have to remove a mirror in RAID10 so that I have room to put in drives that I’m using for initial replication. Here are some other gotchas
The order of drives matters
When changing drives, the Synology DSM actually installs applications and things on a default drive that you set. So when I for instance put a new set of drives in slots 1-6, DSM hung because it couldn’t find the reset of itself.
So what you have to do is put the the “old set of drives” numerically first in bay 1, so that it finds itself. Then you format the “new set as a storage pool”. Then you can reshuffle it back and it all works. So in this sense, there really is a root drive
Shared Folders browsing needs to have the “guest” account active
If you want browsing to work then enable the “guest” account in Control Panel > Users. This is left off as a default and you get all kinds of strange errors. Kodi, for instance, will see the new server, but when you try to browse to it, it will say “unknown error has occurred”. MacOS will show the shares, but when you click, the whole drive will disappear.
Shared Folders in Control Panel and Synology Drive Admin
Shares are created in the Control Panel, not in the File Station application. I constantly find this confusing, but basically, you use Storage Manager to create a Storage Pool and then a Volume in the Pool, then go to Control Panel > File
Finally, if when you want a file share to be used with Synology Drive, so you can sync files and folders to your Mac, you need to go to the server start Synology Drive admin, and enable it. Note it sometimes keeps failing for me so don’t give up and the Enable would keep failing, but eventually, it will work.
Also note that if you have destination Replications of a Shared Folder, then you can’t use Synology Drive with the destination replication which is kind of terrible.
You can hot standby another Synology only if it is of the same type.
There’s a way to do a hot switch with Synology High Availability, but this requires that the two servers be more or less identical in disk array and also model, so pretty useless for home environments. Here a replication seems like a better choice.
Drives are self-labeled and work when moved.
If you pull a drive from another NAS and put it into another Synology NAS, it remembers what it is and what drives form a storage pool. This is awesome in that you can move things around and do copies locally rather than over the Ethernet. I found that with Gigabit Ethernet, I was getting transfers of 30-50MBps, but with an internal transfer it is more like 100-150MBps (about the speed of the hard drives), so this is worth it if you have lots of data. Something that would take a week, takes 1-2 days.
But this label can’t just be erased with Synology. Even if you do a secure erase you will get an error in the Storage Manager. The allocated field, “cannot use with this version of DSM”. You have to pull the drive stick into your PC and reformat it to make it work.
The drives themselves are marked, so moving them to different NASes elements them. This can be confusing for the Storage Pool Manager if you have been doing a lot of moving around. I find that when I have two NASes it’s nice if the physical layouts are the same. If you want to get rid of what it has remembered, make sure drives are *not* in the array, and in Storage Manager, click on the Storage Pool and choose Remove. You will get a scary error message saying it is deleting data (and it will if it is in the array!), but if it is not, it will get rid of the record of it having been in the machine.
Synology DS1812+ latches were not great
The handles and latches on the older DS1812+ snap off after 10 years. Even when I bought new replacements the little butterfly plastic broke off. Not a problem with the new DS2413+ or the RP2423+ which don’t depend on a butterfly mechanism, so progress!
Synology Snapshot Replication Initial copy is great but confusing
To do replication using a disk you will have to find a spare disk. Go to the source NAS and find some room. Hopefully, you have a RAID partition, so just remove a redundant member and put this spare drive in. Then in Storage Manager, create a basic storage pool and create a volume. Then create a Control Panel > Shared Folder, and you create a shared folder (phew!). Finally, go to Synology Snapshot Replication then in replication create a new replication and it will ask you if you want to use a disk for an Initial Copy. Say yes, and point to your scratch drive.
Now pull the scratch drive put it into the Destination. Go to Storage Manager and do an Online Arm which will tell you where the drive came from and remount it.
Click on Snapshot Replication and it will ask for the location of the initial copy, give it the Shared Folder of your scratch disk. It will grind away.
It isn’t very clear to have to go to two places to get the initial copy started. The speed is about 100MBps with a 7200rpm disk. This is probably related to the amount of fragmentation on the drive. Both the source and destination are pretty full. At 80-90%. You should leave HDDs at no more than 80% so YMMV!
At least on our machines, you get a 100MBps copy which isn’t that fast but still 2x faster than Gigabit Ethernet. Of course, if you have say a branch office that is running at 1Mbps, then this is the only practical way to get your data to that office. You create the drive and “sneaker net it over”
Note that in the drive removals, you can theoretically just deactivate the drive and then pull it since all these machines are hot-swappable, but I normally just shut down the machine at home since downtime isn’t an issue.
You have to do this for all the shared folders you want to replicate. If you are lucky they will all fit in your spare drive. If not, then you have to manually delete the shared folder you’ve copied and bring it back to the source and do it over and over until all shared folders are replicated.
In my case, since I already had the data there, it was a pain since really the data is already there, but there is no way using btrfs replication that I could find that would take advantage of this.
