Cache...also in Seafile client!

#1

Hello gentlemen. Here we are with a new problem … indeed … I would say that it is no longer a novelty! This time I am attaching part of the video, but I can’t let you see the whole process as I’m transferring more than 200GB. This should be the scenario:

HDD (236GB of data) -> Upload on Seafile (via Seafile Client)

Now what I say … but because the DATA I have to transfer also passes to the Hard DIsk (I always remember that it is a SSD !!) of the system when they should go directly from the HDD to the seafile server!!! In a nutshell this happens:

HDD (236GB of data) -> Copy data to the OS HDD -> Upload on Seafile (via Seafile Client)

I mean … you realize !? it is a continuous “remove the wax, put the wax”!.. “Remove the data, put the data!”… “Remove Gigabyte, put Gigabyte”!!!
I am attaching a video … which is worth over 1000 words!

In conclusion. I do not recommend this seafile to anyone … if someone looks for suggestions I certainly don’t say “Hey, download seafile”! … I could recommend it ONLY for “SMALL” chores! …
Thank goodness I didn’t buy this software and I stayed at the FREE version! Because that would be wasted money.
I tell you but one thing … the only merit, and i underline UNIQUE, that this software has, is speed … there is no question about this. But as soon as I find a viable alternative, I don’t think twice about it and I run out of this seafile! I’m ruining a PC !!

here is the video

#2

I am not sure I understand what you are saying, but I think I might. Tell me if this isn’t right.

You copy a lot of data into the directory that the client is configured to sync to a library in your seafile server, it notices the new data and starts to work on sending it to the server. While it is working on sending that data you see the seafile client using additional disk space in its cache directory, which you weren’t expecting to see. Is that right?

If so I don’t think what you are seeing is a bug, but is the client working as intended. I will try to explain why that happens. I will warn you that I am not one of the developers, and haven’t read the code, so this is all based on observations and reading the documentation, and it might not be perfectly accurate. Seafile uses a clever de-duplication and sync system that breaks files up into chunks in a consistent way, and then only sends over the network the chunks the other side doesn’t already have (it works both when sending a file to the server, and when the server is sending a file to the client).

This means that if a small change were made to one small part of a large file, only the changed chunks of the file have to be sent, instead of needing to transfer a new copy of the file. This also means that only those small parts of the file that have changed will take additional disk space instead of storing a complete second copy of the file. This makes seafile good at synchronizing the files in a directory with the copy on the server, even over a slow network connection. And this allows the server to have many old versions of files without using many times the space of the original files.

But those advantages do come at a price. The client has to spend CPU time processing files to cut them up into chunks, and it has to cache those chunks while waiting for them to make it over the wire to the server. And it keeps recently used chunks in the cache because if it already has that chunk it doesn’t have to wait for it to come in from the server.

If those disadvantages out weigh the advantages then this might not be the right tool for the job you are trying to use it for.

1 Like