[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Problem with sharing files
Re: Problem with sharing files
Thu, 30 Dec 2021 15:30:14 +0100
thank you for the detailed answer. I will definitely look more into it.
I have tested now again with my laptop using the packages from the AUR.
You were right that I didn't test directly via gnunet-publish from
terminal (only messenger-gtk, cadet-gtk and gnunet-fs-gtk).So gnunet-
publish worked for me now.
However cadet-gtk seems to work fine as well now and even my code in
the current state of messenger-gtk works. ^^'
Maybe it is just a plugin difference or it is a quite unreliable issue
somewhere in the process. I need to check on my main setup again if I'm
able to. But at least I can continue writing the GUI parts for sharing
files in the mean time.
On Sun, 2021-12-26 at 17:50 +0100, Christian Grothoff wrote:
> I looked into this a bit, but could not find it. What is known:
> 1) The client must be doing a datastore PUT or RESERVE operation
> are possible), and expects the datastore to return a STATUS response
> (type 94), but that never happens.
> 2) I checked, and on all code paths the datastore should generate a
> STATUS for PUT/RESERVE requests. So I figured maybe it is blocked on
> some other request, but also that does not seem to be the case from
> code logic. But, hard to be sure.
> 3) Some paths depend on the storage plugin being used (postgres,
> mysql, heap). It might help if you could (a) say which one you used,
> (b) tried with a different storage plugin so we can see if this is
> 4) I tried with 'gnunet-publish' on my system, which worked fine. Did
> you try the CLI, or only gnunet-fs-gtk? It should not make a
> but diagnosing the issue might be easier if we can reproduce it with
> a CLI.
> Last but not least, I did in the past have issues with the database
> it had gotten very big and slow. That's partially because of design
> issues in FS, and partially because of naive usage of the database(s)
> depending on which backend you are using. So another question here
> how big is your database? I'd usually expect it to be empty if this
> you testing things for the first time, but if your peer was running
> a while, it might have grown to many gigabytes and may be really just
> terribly slow...
> My 2 cents
> On 12/22/21 2:55 PM, TheJackiMonster wrote:
> > Hey,
> > I'm currently working on the implementation of file sharing via
> > GNUnet-
> > FS service in the messenger-gtk application. But only the callback
> > from
> > uploading gets called only once with 0/N bytes uploaded and after
> > some
> > time the following warning shows up multiple times:
> > ERROR Request 0x5590f0f01830 of type 94 at head of datastore
> > queue
> > for more than 1 m
> > I've also thought this would be caused by some implementation issue
> > on
> > application side. So I checked using cadet-gtk which worked in the
> > past
> > already (it fails similarly) and even gnunet-fs-gtk provides the
> > same
> > problem. The upload starts but it does not proceed.
> > Maybe someone working on the datastore or FS service knows more
> > details
> > about this or how this could be fixed?
> > Happy hacking
> > Jacki
Description: This is a digitally signed message part