There are now 120 replies in this thread. Iāve tried scrolling around and reading, but itās a lot of noise of āwahh I donāt understand how this concept works and Iām scaredā and ābut MaH FiLeSā, not any actual conversation about the preview build. Some maintenance by OP on the first post would be welcomed.
Looking at OwnCloud, it looks just as bad - a wishy washy āhereās something we cobbled togetherā virtual file system client. I mean, OneDrive has its ups and downs, but itās extremely reliable (it does a hash check & server-side validation before removing any local files, and it only removes local files if I tell it to), and for $10/mo, Iām probably better off, for my sanity, just keeping that.
Maybe Nextcloud will have this feature ironed out next year and I can try again
Stay tuned because after the nextcloud conf we will have some news.
Iām confident that the virtual drive will have some delay/postpone but it will be out for Christmas 2019 i think.
So I have read through the comments here and I explicitely state that I do understand the concept of the virtual drive and the option to force files as offline available in the cache.
Like many others here, Iām very skeptical about the idea of completely replacing folder sync with virtual drives. For me, I really appreciate the existence of a physical copy in the correct directory structure placed on the local hard disk. I can do anything with this directory I can also do with normal directories (because thatās what it is). A few examples:
Even in the case of computer emergencies (e.g. linux single user mode), Iād be able to access the files and copy them to a different device, if needed. I just canāt do that with virtual drives and I loose a lot of flexibility.
I can easily move this directory easily to a different partition in case I run out of disk space.
I can easily spread different synced directories across different local hard disks, just using native file system operations.
@ings Iāve worked with Virtual Drive systems before, and you can replicate what you already have by just āpinningā or āmaking available offlineā your root folder. The Virtual Drive feature is nice if you have Terabytes of data and you only need a few parts locally. If you want everything local like you do now, just make everything āavailable offlineā and it is the same thing as the old client.
@tflidd I tried the Virtual Drive preview for the first time. It errored during install and I wasnāt able to download any files afterwards. =(
@ings I read your message fully, and have used both types of systems and have compared them. I see no difference between designating a directory as āavailable offlineā and the old way of choosing a folder to sync. Can you explain what the difference is in your mind?
Iām still trying to hash out what exactly the VD release means for me too, but the only real difference I can detect is that the the VD version adds the ability to see files/folders that are not available offline.
Isnāt that a rather important difference? I.e, wouldnāt want to have to sync the whole cloud locally. That is rather anti to the idea of cloud storage, isnāt it. Letās hope they can finish the client soon.
@metsuke0 I have made three cases where the virtual disk is not the same as a synced folder:
Computer Emergency, the system doesnāt boot anymore and you only can access the disk from a live system or it is in a limited run mode (e.g. linux single user mode). You might want to copy unsynced work on a usb stick or similar to ensure you donāt loose it. Fact is, in these states the VD client will not be available and you are not having access to your data as stated by the devs above: if the client doesnāt run, the folder is empty / the data is not available. Though the data is somewhere locally stored you simply are not going to find it because the file structure and file names are not the same.
Move a synced directory to a different partition. This is not possible if the VD client doesnāt provide a feature for it. And using symlinks as suggested above is not the same, I want to move the actual data to the partition.
Spread different synced directories across different local hard disks using native file system operations only. Not possible. Only through the VD client if the client might support it.
There are more use cases like above. All of these show that you are completely lost without a running VD client. This is fundamentally different from the situation using synced folders. Accessing a synced folder is easily possible without the sync client running and I see this as a valid and common use case and as a unique feature of synced folders. I donāt argue against virtual drives, there are obvious use cases for these, but I argue against slashing the synced folder functionality.
Nice explanation and I like the option for hiding some of the online folders/files, though that could be challenging if the hierarchy was interrupted with other available folders nested below.
Even for just my personal use, I have so much hierarchy in my folder structure that I donāt want to have to navigate the whole directory to get where I want. Symlinks are a proposed workaround to a new architecture that gives up widely adopted functionality. In my opinion, replacement architecture that requires external workarounds to maintain existing functionality makes a system unreliable for business purposes.
One last point, virtual drives demand hierarchy in a world where, for the past decade, symantic (full text) search tools have been showing that hierarchy is arcane. Is a virtual drive really a step forward or an anchor to the past? Virtual drive platform may offer technical team efficiencies and outsource the sync to other open source teams. But, presumably, that can introduce other risks.
The big reason I have been waiting for virtual drives is due to Excel users. I would like to replace traditional file shares but excel users often times cross link workbooks using network drives.
Yes I could emulate this with folders but the folders would all need to be in the same place on the users computers and be named the same.
Yes i agree that Virtual Drive is a great feature, but itās not enough for deleting the actual behavior of the sync.
As mentioned previously i have two cases where it will be a NON-GO for my clients if nextcloud remove the actual client.
Not all files are text, it helps enormously if these files are associated to projects
Many people still manage their data hierarchically, and for it helps me a lot finding data (itās hard to say if the increased searching time exceeds the time required to store the files)
How do we manage to make a client, that everybody can integrate it in their working routine and that it still is easy to use? There is the virtual FS / complete sync, but as well people who want have one synced folder or seveal synced folders on different locations on their clients (like pictures and documents on different drives). Conflict handling etc.
Nextcloud canāt do a client that fits only your (or my) needs.
@ings
If symlinks wonāt work for you then Iām not sure how to cover points 2 and 3. For more advanced functionality, Iād probably be using a *nix server (as I am now) where Iām more comfortable using the unix tools to do what I need.
During a compute emergency, your data is still on the drive, and the location is not that obscure. You can make a shortcut to the location on your desktop. The VD client even gives you the option to open a folder directly to the data.
@anon9582441 Right, that is the advantage of the VD!
@KarlF12 You still can do the full file sync since VD can make all the files available offline.
@tflidd Let me know if my understanding of the VD system is incorrect. I think I have a grasp on it, but Iām sure youād know better.
Does it have to be in a drive letter? One of the great things about Nextcloud is the ability to sync folders from anywhere on my disks.
For my part, Iād prefer to see the offline files in their usual locations, with placeholders that can be hydrated by the client in those same folders. Sure, if the client isnāt running, the placeholders wonāt exist, but the offline files would be right there in their normal paths. Otherwise, weāre stuck reorganising our content just to fit the client.
IMHO, there are roughly two very different user groups with differing requirements:
PRIVATE USERS, who have their own data storage strategies in place, most of which have grown over many years and are tailored precisely to their individual needs. This data is anchored decentrally, often distributed over several devices.
BUSINESS CUSTOMERS, who (probably) have already a centralized data storage solution in place, collaboration needs and who know and appreciate a file server-like approach (sure, with enhanced funtionality).
In my private environment, where I run a Nextcloud server for family members, relatives and friends, a total of 32 people of different ages and with very divergent technical affinities, the existing Nextcloud client sync concept was understood quite quickly and accepted right away, everyone almost immediately felt at home, nobody asked for a virtual drive funtionality. With the more technically gifted of them I talked about the new sync approach and NO ONE! (except me) found this solution really useful, a few were even overwhelmed with it.
Maybe that my experiences arenāt completely accurate, sort of biased, though certainly they are neither totally exotic. Personally, I think that both approaches have their raison dāĆŖtre and should be available/selectable concurrently. M2C.