Recently Thomas and me met in person and thought about an alternative approach to bring our big file chunking to the next level. “Big file chunking” is ownClouds algorithm to upload huge files to ownCloud with clients.
This is the first of three little blog posts in which we want to present the idea and get your feedback. This is for open discussion, nothing is set in stone so far.
What is the downside of the current approach? Well, the current algorithm needs a lot of distributed knowledge between server and client to work: The naming scheme of the part files, semi secret headers, implicit knowledge. In addition to that, due to the character of the algorithm the server code is too much spread over the whole code base which makes maintaining difficult.
This situation could be improved with the following approach.
To handle chunked uploads, there will be a new WebDAV route, called remote.php/uploads.
All uploads of files larger than the chunk size will go through this route.
In a nutshell, an upload of a big file will happen as parts to a directory under that new route. The client creates it through the new route. This initiates a new upload. If the directory could be created successfully, the client starts to upload chunks of the original file into that directory. The sequence of the chunks is set by the names of the chunk files created in the directory. Once all chunks are uploaded, the client submits a MOVE request the renames the chunk upload directory to the target file.
Here is a pseudo code description of the sequence:
1. Client creates an upload directory with a self choosen name (ideally a numeric upload id):
2. Client sends a chunk:
3. Client repeats 2. until all chunks have successfully been uploaded
4. Client finalizes the upload:
MOVE remote.php/uploads/upload-id /path/to/target-file
5. The MOVE sends the ETag that is supposed to be overwritten in the request header to server. Server returns new ETag and FileID as reply headers of the MOVE.
During the upload, client can retrieve the current state of the upload by a PROPFIND request on the upload directory. The result will be a listing of all chunks that are already available on the server with metadata such as mtime, checksum and size.
If the server decides to remove an upload, ie. because it hasn’t been active for a time, it is free to remove the entire upload directory and return status 404 if a client tries to upload to. Also, a client is allowed to remove the entire upload directory to cancel an upload.
An upload is finalized by the MOVE request. Note that it’s a MOVE of a directory on a single file. This operation is not supported in normal file systems, but we think in this case, it has a nice well descriptive meaning. A MOVE is known as an atomic and fast operation, and that way it should be implemented by the server.
Also note that only with the final MOVE the upload operation is associated with the final destination file. We think that this approach already is a great improvement, because there is always a clear state of the upload with no secret knowledge hidden in the process.
In the next blog I will discuss an extension to this that adds more features to the process.
What do you think so far? Your feedback is appreciated, best on the ownCloud devel mailinglist!
Today, we’re happy to release the best ownCloud Desktop Client ever to our community and users! It is ownCloud Client 1.8.0 and it will push syncing with ownCloud to a new level of performance, stability and convenience.This release brings a new integration into the operating system file manager. With 1.8.0, there is a new context menu that opens a dialog to allow the user to create a public link on a synced file. This link can be forwarded to other users who get access to the file via ownCloud.
Also the clients behavior when syncing files that are opened by other applications on Windows has greatly been improved. The problems with file locking some users saw for example with MS office apps were fixed.
Another area of improvements is again performance. With latest ownCloud servers, the client uses even more parallized requests, now for all kind of operations. Depending on the synced data structure, this can make a huge difference.
All the other changes, improvements and bug-fixes are too hard to count. Finally, this release received around 700 git commits compared to the previous release.
All this is only possible with the powerful and awesome community of ownClouders. We received a lot of very good contributions through the GitHub tracker, which helped us to nail down a lot of issues and improved the client tremendously.
But this time we’d like to specifically point out the code contributions of Alfie “Azelphur” Day and Roeland Jago Douma who contributed significant code bits to the sharing dialog on the client and also some server code.
A great thanks goes out to all of you who helped with this release. It was a great experience again and it is big fun working with you!
We hope you enjoy 1.8.0! Get it from https://owncloud.org/install/#desktop
Often questions come up about the meaning of FileIDs and ETags. Both values are metadata that the ownCloud Server stores for each of the files and directories in the server database. These values are fundamentally important for the integrity of data in the overall system.
Here are some thoughts about what they are why these are so important.This is mainly from a clients point of view, but there are other use cases as well.
ETags are strings that describe exactly one specific version of a file (example: 71a89a94b0846d53c17905a940b1581e).
Whenever the file changes, the ownCloud server will make sure that the ETag of the specific file changes as well. It is not important in which way the ETag changes, it also does not have to be strictly unique, it’s just important that it changes reliably if the file changes for whatever reason. However, ETags should not change if the file was not changed, otherwise the client will download that file again.
In addition to that, The ETags of the parent directories of the file have to change as well, up to the root directory. That way client systems can detect changes that happen somewhere in the file tree. This is in contrast to normal computer file systems where only the modification time of the direct parent of a file is changing.
FileIDs are also strings that are created once at the creation time of the file (example: 00003867ocobzus5kn6s).
But contrary to the ETags, the file IDs should never ever change over the files lifetime. Not on an edit of the file, and also not if the file is renamed or moved. One of the important usages of the FileID is to detect renames and moves of a file on the server.
The FileID is used as an unique key to identify a file. FileIDs need to be unique within one ownCloud, and in inter-owncloud connections, they must be compared together with the ownCloud server instance id.
Also, the FileIDs must never be recycled or reused.
Often ETags and FileIDs are confused with checksums such as MD5 or SHA1 sums over the file content.
Neither ETags nor FileIDs are, even if there are similarities: Especially the ETag can be seen as a checksum over the file content. However, file checksums are way more costly to compute than just a value that only needs to change somehow.
What happens if…?
Let’s make a thought experiment and consider what it would mean especially for sync clients if either fileID or ETag gets lost from the servers database.
If ETags are lost, clients loose the ability to decide if files have changed since the last time that was checked by the clients. So what happens is that the client will download the files again, byte-wise compare them to the local file and use the server file if the files differ. A conflict file will be created. Because the ETag was lost, the server will create new ETags on download. This could be improved by the server creating more predictable ETags based on the storage backends capabilities.
If the ETags are changed without reason, for example because a backup was played back on the server, the clients will consider the ones with changed ETags as changed and redownload them. Conflict handling will happen as described if there was a local change as well.
For the user, this means a lot of unnecessary downloads as well as potential conflicts. However, there will not be data loss.
If FileIDs got lost or changed, the problem is that renames or moves on server side can no longer be detected. That would result in a new download of files in the good case. If a fileID however changes to something that was used before, that can result in a rename that overwrites an unrelated file. That is because clients might still have the FileID associated with another file.
Hopefully this little post explains the importance of the additional metadata that we maintain in ownCloud.
Incremental Sync is probably the feature that most people ask, or even sometimes cry for. Recently there was another wave of discussion about ownCloud is doing incremental sync or not.
I will try again (as in this issue) to explain why we decided to slowing that feature. Slowing means that it will be done later, not never, as it was stated. It is just that we think that other things benefit the whole idea of ownCloud more. That has plain technical reasons. Let’s dive a bit into.
RSync is great
Nobody will object here. In a nutshell, this is how rsync works: There is a file on the client and on the server. The idea is to not transfer the entire file from one side to the other if either side changes, but only the parts that have changed.
The original rsync does that by chopping the file to blocks of a given size and calculating a checksum of each of the blocks. The list of checksums is sent to the server and – here’s the trick – the server looks at its version of the file and for each of the checksum in the list, it seeks if it finds the same block in the file. That will often not be at the same position in the file, but maybe somewhere else. That is done for each block, and finally the server will work out the information of which parts of the file are existing and which are not and have to be sent by the client.
By way of this clever algorithm, we will just have to transmit a very small fraction of the changed file, because most content did not change. And that is what we want! Yeah!
Mission accomplished? No, not really. While there is basically nothing wrong with the idea in general, there is a severe architectural downside. The rsync algorithm depends on a strong server component which, for each file, searches around and calculates checksums. In an environment where we potentially have a lot of clients connecting to one server that would create a huge load on it which we need to avoid. So what if instead of putting the burden on the server’s shoulder, we could make the clients take the responsibility?
And guess what, there has been somebody thinking about that before and he says:
Use ZSync for this!
ZSync basically turns the idea of rsync upside down and shifts the calculation of checksums away from the server and onto the clients. That means that with zsync, the server can keep a static list of checksums for every block specific to a version of a file. The list can for example be computed along the upload of the file to the server. From that point it does not change, as long as the file does not change. That means less computation work for the server, and maybe this job can also put into the client.
So far that sounds cool (even though some questions remain) and sounds like something that can help us.
Unfortunately, the approach does not work very well for compressed files. The reason is that if a file gets compressed, even if only a couple of bytes in the original file change, the compression algorithm usually changes a lot all over the entire file. As a result, the zsync algorithm can only compute a comparably large diff. Given the cost of computation that can turn inefficient quickly.
“But who uses compressed files?” you might argue. The problem is that almost each and every of the files in everyday life are stored compressed. This is for example true for Microsoft Office files and the Open Document files produced by LibreOffice and Apache OpenOffice. They are really renamed ZIP containers, that hold the documents with all its embedded files, etc.
Now of course you will reply that zsync has an improved algorithm for compressed files. Yes, true, that is a great thing. However, it involves that the compressed file gets uncompressed to be worked on by zsync. Afterwards it is compressed again. And that is the problem: As common compressors do not leave a hint behind _how_ the file was compressed, it is not possible to reliably recreate a file that is equivalent to the original one. How will apps react on a file that has changed its compression scheme?
As said above, yes, we will at one point of time implement something along the zsync algorithm. The explanations above should show however, that at the current state of ownCloud, other features will improve ownClouds performance, stability and convenience more. And that is the important thing for us, more than pleasing the loudest barking dogs.
Here is a rough outline of how I would move on on this, open for your suggestions and critique:
The zsync algorithm is designed to improve downloads. We need it for both up- and downloads, and it needs to be thought through if that is also possible. For the server side functionality, there are a couple of open questions which carefully have to be investigated. Preferably an app can be written that provides the handling of the zsync checksum lists. That has to be clarified and discussed, and that will take a while.
But as outlined above, this idea is only clever for a limited amount of file types. So what I would suggest first is that we get an idea of the file types users usually store in their ownCloud, so that we can do a validated estimate on how this feature helps. I will follow up on this first step.
Thanks for reading this long blog post. Thanks Danimo for lectorate.
Our recent ownCloud Client 1.7.0 release contains the new feature of overlay icons in GNOME nautilus, MacOSX and Windows. That is nice, but that makes us as old KDE guys sad as Dolphin was missing on the list.
That needs to change, and here we go: Olivier Goffart wrote a patch to do overlay icons also in Dolphin, which was not straightforward, because in addition to an dolphin plugin, also a patch for libkonq was required.
We prepared some test packages in our development repository isv:ownCloud:devel for those who wanna try and know their way around. Current it only builds for a couple of openSUSE Distros. You need to install kdebase4 and dolphin-plugins and after installation, it’s easiest to restart KDE to make it registered. But be warned: The two packages replace packages from the previous installation, only do it if you really know what you’re doing!
It would be great if at least the libkonq patch could make it to upstream, and I would appreciate if somebody who is a bit more fluent with recent KDE libs development could give me a hand on that. Otherwise, if distros wanna pick up the patches to make the overlays work, of course the patches are here: patch for libkonq and the ownCloud Dolphin plugin. The plugin will work with the released version 1.7.0 of ownCloud Client.
Last week, Thomas, Christian and myself were attending a workshop in CERN, the European Organization for Nuclear Research in Geneve, Switzerland.
CERN is a very inspiring place, attracting intelligent people from all over the world to get behind the secrets of our being. I felt honored to be at the place where for example the world wide web was invented.
The event was called Workshop on Cloud Services for File Synchronisation and Sharing and was hosted by CERN IT department. There have been around 100 attendees.
I was giving a talk called The File Sync Algorithm of the ownCloud Desktop Clients, which was very well received. If you happen to be interested in the sync algorithm we’re using, the slides are a nice starting point.
What amazed me most was the great atmosphere and the very positive attitude towards ownCloud. Many representatives of edu organizations that use ownCloud to which I talked were very happy with the product (even though there are problems here and there) from the technical POV. A lot of interesting setups and environments were explained and also showcased ownCloud’s flexibility to integrate into existing structures.
What also was pointed out by the attendees of the workshop was the importance of the fact that ownCloud is open source. Non free software does not have a chance at all in that market. That was the very clear statement in the final discussion session of the workshop.
The keynote was given by Prof. Benjamin Pierce from Pennsylvania with the title Principles of Synchronization. He is the lead author of
the project Unison which is another opensource sync project. It’s sync engine marks very high quality, but is not “up-to-date software” any more as he said.
I had the pleasure to spend quite some time with him to discuss syncing in general and our sync algorithms in particular, amongst other interesting things.As part of his work, he works with a tool called QuickCheck to do very enhanced testing. One night we were sitting in the cantina there hacking to adopt the testing to ownCloud client and server. The first results were very promising, for example we revealed a “problem” in our sync core that I knew of, which formally is a sync error, yet very very unlikely to happen and thus accepted for the sake of an easier algorithm. It was impressive how fast the testing method was identifying that problem.
I like to follow up with the testing method.
Furthermore we met with a whole variety of other interesting people, backend developers, operators of the huge datasets (100 Peta-Byte), the director of CERN IT, a maintainer of the Scientific Linux and others.
Also we had the chance to visit the Atlas experiment, it is 100 meter underneath the surface and huge. That is where the particles are accelerated, and it was great to have the chance to visit that.
The trip was a great experience and very motivating for me, and I think it should be for all of us all doing ownCloud. Frank was really hitting a nerv when he was seeding the idea, and we all were doing a nice product of it so far.
Lets do more of this cool stuff!
Yesterday we released ownCloud Client 1.7.0. It is available via ownCloud’s website. This client release marks the next big step in open source file synchronization technology and I am very happy that it is out now.
The new release brings two lighthouse features which I’ll briefly describe here.
For the first time, this release has a feature that lives kind of outside the ownCloud desktop client program. That nicely shows that syncing is not only a functionality living in one single app, but a deeply integrated system add-on that affects various levels of desktop computing.
Here we’re talking about overlay icons which are displayed in the popular file managers on the supported desktop platforms. The overlay icons are little additional icons that stick on top of the normal file icons in the file manager, like the little green circles with the checkmark on the screenshot.
The overlays visualize the sync state of each file or directory: The most usual case that a file is in sync between server and client is shown as a green checkmark, all good, that is what you expect. Files in the process of syncing are marked with a blue spinning icon. Files which are excluded from syncing show a yellow exclamation mark icon. And errors are marked by a red sign.
What comes along simple and informative for the user requires quite some magic behind the curtain. I promise to write more about that in another blog post soon.
Another new thing in 1.7.0 is the selective sync.
In ownCloud client it was always possible to have more than one sync connection. Using that, users do not have to sync their entire server data to one local directory as with many other sync solutions. A more fine granular approach is possible here with ownCloud.
For example, mp3’s from the Music dir on the ownCloud go to the media directory locally. Digital images which are downloaded from the camera to the “photos” dir on the laptop are synced through a second sync connection to the server photo directory. All the other stuff that appears to be on the server is not automatically synced to the laptop which keeps it organized and the laptop harddisk relaxed.
While this is of course still possible we added another level of organization to the syncing. Within existing sync connections now certain directories can be excluded and their data is not synced to the client device. This way big amounts of data can be easier organized depending on the demands of the target device.
To set this up, check out for the button Choose what to Sync on the Account page. It opens the little dialog to deselect directories from the server tree. Note that if you deselect a directory, it is removed locally, but not on the server.
There is way more we put into this release: A huge amount of bug fixes and detail improvements went in. Fixes for all parts of the application: Performance (such as database access improvements), GUI (such as detail improvements for the progress display), around the overall processing (like how network timeouts are handled) and the distribution of the applications (MacOSX installer and icons), just to name a few examples. Also a lot of effort went into the sync core where many nifty edge cases were analyzed and better handled.
Between version 1.6.2 and the 1.7.0 release more than 850 commits from 15 different authors were pushed into the git repository (1.6.3 and 1.6.4 were continued in the 1.6 branch which commits are also in the 1.7 branch). A big part of these are bug fixes.
Who is it?
Who does all this? Well, there are a couple of brave coders funded by the ownCloud company working on the client. And we do our share, but not everything. Also coding is only one thing. If you for example take some time and read around in the client github repo it becomes clear that there are so many people around who contribute: Reporting bugs, testing again and again, answering silly looking questions, proposing and discussing improvements and all that (yes, and finally coding too). That is really a huge block, honestly.
Even if it sometimes becomes a bit heated, because we can not do everything fast enough, that still is motivating. Because what does that mean? People care! For the idea, for the project, for the stuff we do. How cool is that? Thank you!
Have fun with 1.7.0!