There was a server config issue causing some folks to get full-page 403 errors on visiting toot.cat.
The problem seems to be limited to visitors accessing the site via http without having recently accessed it via https, e.g. if you just typed "toot.cat" in the URL bar.
That's been fixed now.
I've also been seeing sporadic 50x errors pop up inside the Mastodon UI; that seems probably unrelated.
As always, please let me know if you know of any issues using #tootcat.
After some hunting around, I finally figured out that I needed to bump a couple of parameters *way* up -- like, from 25 to 100.
Latency is now back in reasonable territory and outgoing toots should be going through in a matter of seconds or single-digit minutes, not hours.
Let me know if this is not the case.
Technical stuff here: https://htyp.org/User:Woozle/toot.cat/2018/06/22
Rebooting #tootcat to apply a config change and also because the server keeps asking me please can I reboot it because of a kernel upgrade a little while ago.
Normally I'd just reload or restart the daemon or whatever.
Fasten your seatbelts...
So, umm, sorry about that SSL outage that maybe you didn't notice.
It turned out to be an IPv6 problem.
I've had this exact same problem so many times, I don't know why it took me 4 hours to figure it out.
Very tired. All for now.
I think I asked before, didn't get any responses, and then didn't do anything.
A bit more time has passed, this account is a bit more connected to the actual #tootcat usership, so I think this time I might actually do it..
Does anyone have any objections to one of these becoming the site's favicon? (It would replace the standard Mastodon m-fist.)
Any prefs of one over the other? Comments?
I think I kind of like the wider one, but it might not work as well at icon size.
GitLab open sourced push mirroring and added it core. This is a great win for #decentralization!!! You can now run your own canonical GitLab instance & easily mirror to GitHub for findability.
Read more here: https://about.gitlab.com/2018/05/22/gitlab-10-8-released/#push-mirroring-now-open-source
Another reason why GitLab is #1
#tootcat founder @polymerwitch has returned as a moderator to help out with the spambots we've been dealing with recently.
General technical issues should still be reported primarily to @woozle , with @LottieVixen as backup.
Welcome back to the team, Briar Rose! 🌹
Spammers may now witness the firepower of this fully-armed and operational instance.

In with all the spambots I see a new account called @townler, apparently intended as social media presence for an Australian dating site.
At minimum, they should agree to link to their toot.cat account on their home page and toot something about what drew them to this site -- just so I know they're not apamming every instance they can manage...
...but I'm more than halfway inclined to suspend without further review, even though commercial content is not against our current CoC.
Thoughts?
More new accounts silenced; alert @woozle if you were silenced in error.
Also, don't worry about spamming us with reports -- it saves work on our end.
Thanks again for your patience while we weather this annoyance ^.^
Any new accounts that posted spam have now been suspended, and new accounts that haven't introduced themselves (<waves hi to zac>) have been silenced.
If you have been silenced but are in fact not a bot, please ask @woozle to un-silence you.
We'll continue silencing new accounts that follow the patterns I've noticed.
Everyone else: please do report any spam you're still seeing.
Thanks, everyone, and sorry for the nuisance!
NOTICE TO RECENT SIGNUPS: I'm going to give new accounts a minimum of 24 hours to show some signs of not being bots, and then suspend any that seem unlikely to be anything else.
So, as I said earlier: at least say hi or something. Give me something checkable in your profile. Do something non-botlike. If you're not sure, ask me and I'll personally give you a short Turing test.
Thanks again.
I'm noticing a suspiciously large number of new signups with names that look like they were generated by an algorithm.
Can the new folks within the past 48 hours or so please introduce themselves? Thanks!
Post-mortem on the reboot issue is here: https://htyp.org/User:Woozle/toot.cat/2018/05/20
We're back, I think... need to see how federation is doing.
Tentatively, the problem isn't on our server -- but I'm going to go ahead and reboot now, just to get that out of the way.
If there are problems, I'll post updates on Mew (which runs on a different server): https://mew.toot.cat/mw/Main_Page
We seem to be having a blockage on outgoing toots. I'm hoping that rebooting the server (which it's needed for weeks anyway) will fix the issue, but I'm physically in a place where it's difficult to focus and I don't want to do a reboot until I'm prepared to deal with any squidgy Masto issues that might show up.
I expect I'll be rebooting about 1-2 hours from now.
#tootcat EMAIL PROBLEM FIXED i think!
Installed Postfix on local server, reconfigured Mastodon to use that... and, oh yeah, you have to restart the SideKiq service, not just the web service.
I suppose this should have been obvious, but also nobody mentioned it ever.
And contrary to widespread belief, setting up Postfix is easy. <click>
Please let me know if you continue to experience any issues with email notifications.
-- your friendly TootCat Admin Team with a headache (no extra charge)
The #tootcat server wants a reboot, because of system package updates.
I'll try to do it early in the morning, and will post a message here first.
cc: @LottieVixen
We're experiencing an outgoing email outage on #tootcat. I don't know what changed, but the problem seems to be that our relay server (on a different machine) isn't accepting IPv6 connections.
This is fixable, once I clear out a couple of other issues standing in the way.
Here's the trouble-ticket, if you want to keep an eye on it: https://rm.vbz.net/issues/97
We were experiencing delays of up to 3-4 hours with toots from #tootcat being received by other instances.
We've upped concurrency on SideKiq from 5 threads to 25, which is rapidly reducing the backlog of "push" requests, which we sincerely hope is what's causing the problem.
It looks like this should be cleared up within an hour or so.