<DasSkelett[m]>
Did you check the gunicorn instances? My theory is still that one or more of them is not up, and whenever the web server delegates a request to those instances it fails.
<DasSkelett[m]>
Could be gunicorn automatically restarting when it fails to come up?
<DasSkelett[m]>
That's the best guess I can take without looking at logs and stuff myself.
<DasSkelett[m]>
Hm, something like graylog?
<DasSkelett[m]>
gunicorn instances failing to start should be visible in syslog.
DasSkelett[m] has quit [Excess Flood]
Vali[m] has joined #spacedock
<Vali[m]>
ping 🏓
HebaruSan has joined #spacedock
Vali[m] has quit [Client Quit]
VITAS has joined #spacedock
VITAS[m] has joined #spacedock
VITAS[m] has quit [Excess Flood]
HebaruSan[m] has joined #spacedock
<HebaruSan[m]>
(before the fix)
<HebaruSan[m]>
Well that's not something I ever would have guessed 🙂
<HebaruSan[m]>
Nice, failsafe!
<HebaruSan[m]>
Ahh, that makes sense, thanks. And it meant that some percentage of CKAN requests happened in the middle of a restart, and failed.
<HebaruSan[m]>
Any other consequences of extended celery downtime? Mod scores not getting set, etc.?
HebaruSan[m] has quit [Excess Flood]
M_discord_652889938043338755[m has joined #spacedock
<M_discord_652889938043338755[m>
Ok
<M_discord_652889938043338755[m>
this is the server ?
<DasSkelett[m]>
See the code I linked, it explicitly redirects to `"http://" + cdn_domain + '/' + mod_version.download_path`.
<DasSkelett[m]>
Or just use the already existing `protocol` config option, and assume that whoever sets up another instance of SpaceDock and makes use of a "CDN" will make sure that that `cdn-domain` supports the same protocol as the main site itself.
<DasSkelett[m]>
Yeah in this case I think I accused them unjustly. I thought they introduced this for all sites not on some whitelist or something. But a warning on HTTPS websites that (accidentally or not) offer downloads via HTTP makes sense.
DasSkelett[m] has quit [Excess Flood]
<DasSkelett[m]>
We should change that to `https://`. Or maybe add the scheme to `cdn-domain` in the config, so it's configurable.
M_discord_652889938043338755[m has quit [Client Quit]
<VITAS[m]>
so bottom line bta has something ready to deploy?
<VITAS[m]>
beta
<VITAS[m]>
if yes: 8am cest is the time with the lowest hit count so if DasSkelett and you are arround tomorrow at that time i would do the deployment then.
<VITAS[m]>
and it consists of fixes for both thumbnail and dl problem?
<VITAS[m]>
and thx for the work :)
<VITAS[m]>
how well are thos other fixes tested?
<VITAS[m]>
maybe just sleep all the time
<VITAS[m]>
not if swatch beats finaly take over :P
<VITAS[m]>
:D
<VITAS[m]>
but imagine all those fashon victims, hippsters and youtube people that dont know what to do now!
<VITAS[m]>
Btw theres a CKAN question in SD forum thread you might want to look at.
<VITAS[m]>
just heared the fashon industry is in crisis becasuse people dont want fancy cloth anymore just sweat pants
<VITAS[m]>
i hate that branding
<VITAS[m]>
mental horizons as far as MEEEEE!
<VITAS[m]>
how do you zoom from your iphone? twitter me!
BarronTrump[m] has joined #spacedock
RockyTV[m] has joined #spacedock
<RockyTV[m]>
hm, irc messages are sent here but not the other way around (discord/matrix messages aren't sent to irc)
<RockyTV[m]>
okay
<RockyTV[m]>
> has quit (Excess Flood)
<RockyTV[m]>
discord-irc bridge is down btw
<RockyTV[m]>
is there a way to prevent the bridge from sending old messages?
DasSkelett[m] has joined #spacedock
HebaruSan[m] has joined #spacedock
<VITAS[m]>
pling
<VITAS[m]>
thx fixed
Github[m] has quit [Quit: Client limit exceeded: 6]
mastops12[m] has joined #spacedock
<VITAS[m]>
?
BarronTrump[m] has quit [Quit: Idle timeout reached: 10800s]
RockyTV[m] has quit [Quit: Idle timeout reached: 10800s]
DasSkelett[m] has quit [Quit: Idle timeout reached: 10800s]
HebaruSan[m] has quit [Quit: Idle timeout reached: 10800s]
Zack3000000000[m] has joined #spacedock
<Zack3000000000[m]>
hello?
<Zack3000000000[m]>
hello?
<VITAS[m]>
hello?
<VITAS[m]>
hello?
<Zack3000000000[m]>
does anyone know how to install the Parallax mod for ksp?
<VITAS[m]>
i guess the mod author does
<VITAS[m]>
have you tried contacting him on the forums?
mastops12[m] has quit [Quit: Client limit exceeded: 6]
VITAS[m] has quit [Quit: Client limit exceeded: 6]
DasSkelett[m] has joined #spacedock
<DasSkelett[m]>
Alternatively, scroll up, explained the whole process above already <span class="d-mention d-user">Zack3000000000</span>
<Zack3000000000[m]>
ok one sec
<HebaruSan[m]>
Don't mind me expressing some mild annoyance there at the end of the thread
<HebaruSan[m]>
The mod banner is wider than most people's monitors
<DasSkelett[m]>
Wow, that image is really big.
<HebaruSan[m]>
And it takes 2-3 seconds to load on my network
Zack3000000000[m] has quit [Quit: Client limit exceeded: 6]
VITAS[m] has joined #spacedock
<VITAS[m]>
And thats how large mods come into existance :D
<VITAS[m]>
btw make sure that everything that needs to be deployed tomorrow morning is in the right branch and known to work :)
<HebaruSan[m]>
Which is the right branch in your view of things? Right now we have all the desired changes advanced up to beta, but not yet master
<VITAS[m]>
master i guess
<HebaruSan[m]>
OK, when would you want that? I like to minimize the time that master differs from the production server
<VITAS[m]>
i wrote on the forum 8am cest on the 5th (tomorrow)
<VITAS[m]>
all we need to deploy is the chrome/dl fix and idialy the thumbnail fix
<VITAS[m]>
the rest dont have to be deployed yet if you dont feel comftable
<HebaruSan[m]>
So 11.5 hours from now?
<VITAS[m]>
yes because chrome fix is a race against time
<HebaruSan[m]>
OK, sounds good, we'll do some additional testing and merge PR 324 at some point before then
<VITAS[m]>
cool thx :)
<DasSkelett[m]>
So far I've tested mod creation, update, download, notifications, locking, unlocking, download stats, header images and thumbnails.
<VITAS[m]>
do you want to put all thes on prod?
<VITAS[m]>
(tomorrow)
<DasSkelett[m]>
That's just the SpaceDock features I tested, independent of particular pull requests/changes.
<DasSkelett[m]>
To confirm we didn't break anything on the way.
<VITAS[m]>
ah you tested those pages with the fix for mixed cprotocols
<DasSkelett[m]>
And yes, we should deploy all changes that are in beta. Otherwise our branches will get really really messy.
<VITAS[m]>
if you think they are good to go ill do it
<HebaruSan[m]>
Agreed, not just because of branches, but because the current changes are good to have. Not just thumbnails and HTTPS, but also things like the ATS purging fix.
<HebaruSan[m]>
And most have been waiting in alpha/beta for a while now, they're not going to get any more tested just sitting there without a deadline
<VITAS[m]>
if it works in prod i fully agree that they are nice to have :)
<VITAS[m]>
i want to minimize the posibillity of unforseen problems when deploying
<DasSkelett[m]>
We can't know if they work in production until they are in production 🤷♂️
<VITAS[m]>
yes what we can do is deploy it ont he standby prod and see how it fairs with prod data
<VITAS[m]>
we cant alter any data but we can login and dl mods and so on
<DasSkelett[m]>
Well we _do_ need to apply the migrations though, otherwise we can't test anything because the backend would spam hundreds of database errors.
<VITAS[m]>
that again so we have to wait then :/
<DasSkelett[m]>
You can copy production db if you want. But that's how it is.
<DasSkelett[m]>
Ugh, found a small visual issue. Another side effect of #293 fixing the `bootstrap-theme.css`. How could someone back then think it would be a good idea to edit this file directly?
HebaruSan[m] has quit [Quit: Client limit exceeded: 6]
<VITAS[m]>
i did think it was the only way to overwrite the direct edeting and wrong class application taken over from ks
<DasSkelett[m]>
Embarassing that our site only works with broken CSS syntax.
<VITAS[m]>
yes
<VITAS[m]>
its a mess
<VITAS[m]>
for years
<VITAS[m]>
all ive gotten is attempts by people to start froms cratch
<VITAS[m]>
and they lost interest 10% trough
<VITAS[m]>
so unless someone does areal attempt at redoing the theme in current bootstrap without lofty goals we wont have a pridictable frontend
<HebaruSan[m]>
<span class="d-mention d-user">DasSkelett</span> would you prefer to create the PR for `service-icon` or review it?
DasSkelett[m] has quit [Quit: Client limit exceeded: 6]
HebaruSan[m] has quit [Quit: Client limit exceeded: 6]
DasSkelett[m] has joined #spacedock
<DasSkelett[m]>
Works either way for me. Confirmed that disabling/removing the `background-color` property fixes it in the Firefox dev tools.
HebaruSan[m] has joined #spacedock
<HebaruSan[m]>
Is `service-icon` a true part of bootstrap or a KerbalStuff fabrication?
<DasSkelett[m]>
I think it's KerbalStuff customisation, the way it looks. But hard to say.
<VITAS[m]>
there are icon names as part of the iconset
<VITAS[m]>
there isnt a service-icon in bootstrap
<DasSkelett[m]>
Then we could also think about moving the whole block into `index.scss`, we aren't using `service-icon` anywhere else. But I'm fine with keeping the changes minimal for now too.
<HebaruSan[m]>
Yeah, I've done that locally, looks fine for now
<VITAS[m]>
i hopy you guys are there when i deploy that stuff