Zum Inhalt der Seite gehen


@Friendica Support
In the last time I have really problems that all post arrive at my node (or at least are visible in the correct way). Now it happens that my comment on a post of @Tobias didn't arrive his node.
It seems that my link to the source is wrong (My wrong source link https://social.diekershoff.de/display/f3ad7b1c6557827cfe8e738613202506 isn't available on his node).
Yesterday I wrote a PM to @Tobias with request to answer if he got the PM. Since he haven't answered I guess he didn't get the PM.

@Tobias
I tried to comment to the post where you mentioned the css issue.
Here my reply:
No. It seems to be a bug in friendica. But it seems that this bug doesn't have such a bug visual impact in other themes. Have a look at https://friendica.kommune4.de/display/4029d2ca665766005cb0fcd608950607
The post with the guid f3ad7b1c6557827cfe8e738613202506 is available on my system. But since it isn't public, I guess you can't see it on Tobias' system. But I don't see a comment there. So it hasn't arrived there.

I also had posting problems (with @Ben Utzer - but I cannot reproduce it by now on my system, so I can't tell where the problem is.
I will have a look for that information you asked for. But in that case it also was a non public post.
The posting is there, but it is a private posting, so if you are not authed on my system it tells you that the posting is not there at all.

And you are right about the PM it did not reach me.

I've never encountered such a thing before in quattro or clean. Since you did some experimental things in frio I suspected it being a bug in frio. But Mike seemed to be using vier at that time.
It looks like /poco is returning 401s from lighttpd
and /poco/tobias gives a 404, using the lighttpd config rewrite rules from the example config in the repository.
Though, thinking about it I don't share my contacts, POCO was disabled then right?
I just got this message right now, so I assume deactivating the Retriever might have helped to receive this comment. See https://github.com/friendica/friendica/issues/2675

Not sure, but @Michael Vogel can probably answer that question if needed.
The posting is there, but it is a private posting, so if you are not authed on my system it tells you that the posting is not there at all.

But when I visit your profile page I should be authenticated because we have a duplex friendship
I'd rather say my system took a very long time working on the queue.

We had a micro-power outage in the morning, after my postings. After the reboot the system was rather unhappy I'd say, looking at the load history in monitorix, there where several high load occasions far above the limit for background processes. The time is not accurate in the plot, but it might match the queue delivering many postings, spawning to many processes before the load limit was reached.

Or do delivery processes from the queue count as foreground process @Michael Vogel then I need to lower that number...
Yes, you should and I am authentificated when following a magic link to your profile.
As said above, the only thing that is logged with an error from kommune4.de I found is the POCO stuff, but since I have my contact list private, I think that is ok.
The poco stuff doesn't matter.
So if I click on the posting time in Vier that should happen?
You need a magic link, on most themes the icon changes, the URL is yournode.tld/redir/123 or some other number. When following it you should land on the node of your contact and be creeted by a popup saying Hello. And you should get the possibility to comment on postings.
yes, I realized that later over the day too. But that's the only error in the access log, except calls to the old subdomain.
@Tobias I just tried to update your contact (contact edit / update) and I get this in the logs
2016-07-13 08:51:46@log57860122b36dd1.68129519  [NORMAL]:onepoll.php:43:onepoll_run     onepoll: start

2016-07-13 08:51:48@log57860122b36dd1.68129519  [NORMAL]:onepoll.php:152:onepoll_run    onepoll: poll: (298) IMPORTER: Rabuzarus, CONTACT: Tobias

2016-07-13 08:52:07@log57860137e1ab50.95606945  [NORMAL]:network.php:361:http_status_exit       http_status_exit 404
Ill check the logs later. ..
That was a request to /statistics.json IIRC I have not activated the export of statistics to th federation info overview. .. @Michael Vogel should know if it then hidden or we have to check the rewrite rules for lighttod from the samples.
hmm, we should have a better logging description. so we what part gets a http_status_exit
The check for the statistics is optional. When a contact is added or changed then always a gcontact is added or changed as well. And when this is done then the server is checked as well.
We have an enhanced logging since some versions. Look at the "log57860122b36dd1.68129519" in the data above. The 404 is generated by the process "log57860137e1ab50.95606945" which is an independent process from "log57860122b36dd1.68129519" which does the probing.
ahh, thanks for clarification
Checking the different end points are different steps right? Could not the entry with the 404 mention the endpoint checked?

Saves a grep over the file with the logid
This could be anything - but nothing in the probing process because the whole check process is a single one.
Ah, ok - thanks for clarifying it. Good to know about the logging IDs :-)