Skip to main content


 
!iSurf Support

Today I noticed I cannot type in the reply box. I have to click comment on someone else's comment to post to an existing thread. Did something change on the server?

Without doing anything in Chrome, it suddenly works (been out sick).

@Shelenn Ayres Cache probably updated. Seems to be fixed.



 

Missing in global directory


!iSurf Support

When does a new account generally become listed in the global directory? #newhere and thanks! Loving Friendica so far. I did try searching....

Hello @Brian Millan I guess you've properly set to be listed in the global directory (GD). I think there are a range of factors that contribute to this problem.

Your node has quite a lot of users but is geographically far away in the network from the GD.

I'm hosting one of the global directories: https://dir.hubup.pro/. I tried to manually increase the pulling interval from this GD (located in Southeast Asia) of your node (in North America). I noticed that most of the time, the pull is not successful and on the few occasions when it was successful it pulled only three profiles as specified by the current code.

I can see in my GD's database your own profile URL being scheduled for pulling (that means getting the relevant public information from your profile). But by the current rate of success it may take a very long time for new profiles on your node to be listed.

... show more

Thanks Andy, I see now. Distributed, decentralized systems are quite interdependent on each other and net conditions aren't they? (rhetorical). Of course this is all fun to me and I was interested to find out the reasoning. Adam said to me at one point, don't worry, it will find it's way in eventually. Thanks again!
@Andy H3
This entry was edited (2 months ago)

@Brian Millan actually we'd like to thank you for highlighting this issue by persistently following up on this. I'll be speaking to the developers of the GD code and will use your case to show the current shortfall in the current design. By the way, you're now listed in this global directory at least.. https://dir.hubup.pro/search?q=Brian+Millan

Hey @Andy H3 , thanks for following up on things this weekend.

This was an item on my to-do list to follow-up on, as my node reports to my own GD at https://socialdir.isurf.ca and @Brian Millan hadn't shown up last time I checked, which was at least a couple of weeks after I mentioned it should show up eventually. I didn't expect it to take this long, however.

Perhaps we could open up a thread somewhere for a separate discussion about the issues you've had polling some of that information from my system. I'd like to know what kind of errors have been reported and why the hit rate would be so low. It would be something I would like to help see sorted out.

Hi @Adam Clark I think we should open an issue here: https://github.com/friendica/friendica-directory/issues

I run this console command:
bin/console server-poll [url=https://socialdir.isurf.ca]https://socialdir.isurf.ca[/url]
and ended up mostly getting "unsuccessful" responses (despite seeing your sever online) .

I was thinking about increasing the timeout settings, but I'm not sure yet if the source code actually permits this. I'd like to run some more tests and see how the poll responses compared to ping values.

One thing that is note worthy is that your GD in California [?] seems to experience a similar issue than my GD in Thailand or the main one in France, when polling from the isurf node in Canada . Definitely something we need to look it.

There's definitely something odd with that 'server-poll' console command itself, as it seems to bail on any error even though it's queued up a lot of work to do. So rather than skipping problems and moving on with the rest of it's work, it just seems to give up.

However, is that the same mechanism that is used for getting the profiles from the friendica nodes initially? I'm not sure what process is used to get the directory aware of a profile in the first place.



 
!iSurf Support

Question: I've been experimenting with features. On the protocols box in the left panel, When I choose DFRN I see user posts showing from Friendica. When I choose Activity Pub I see user posts showing from Friendica (AP). I've seen these in mixed form (both) from this node. What does it mean?

Prior to the 2019.01 stable release of Friendica, DFRN was the only way Friendica talked to other Friendica nodes. I'm not sure if it was used beyond Friendica - perhaps Hubzilla as well.

With the introduction of AP, Friendica is migrating away from DFRN and the system tries to figure out which protocol it can use, preferring AP. I suspect Friendica will have to support DFRN for some time, so that nodes running older versions will still be able to federate, but the new releases I believe will continue to have more and more AP support and prefer to use it when possible.

So, right now it's a mix with Friendica nodes, and I couldn't tell you all the ins and outs of what magic it's doing to figure out which protocol it should use.

Also, please forgive me where I've stated things as absolute fact - I believe I know the details only to a degree, I could be off about some of it. The gist of it is there, and there's talk of it on the prior Friendica release announcement.

HTH!



 

Google+ connections may stop working as of today


!iSurf Support

Hi all,

As you probably have already seen, some parts of Google+ are beginning to shut down.

I have received a couple of email notifications from Google about their API being shut down as of today, which means that posts to/from Google+ from here will stop functioning.

Some people are apparently still seeing some functionality so I suppose it's possible that it could take some time for this feature to stop operating completely, but it's probably good to know that it will be stopping very soon.

Apparently the shutdown is not orderly and all users are not affected by different things stopping at the same time. LOL

What I find most amusing with Google's emails on the subject is they state: "This has
been a progressive shutdown where calls to affected APIs began
intermittently failing on January 28, 2019."

That may be true, but aside from problems with our own settings locally, the API seems to have been working pretty consistently well.