Feedback Technical Grievances

[WolfeTone]

Oh fuck oh shidd what is this I'm not good at computer

 

[Null]

So is the weird formatting stuff because of the recent introduction of Kiwiflare? It seemed fine the past few weeks.

• There is an issue where blank pages are being sent randomly. This also affects fonts, styles, images, and JavaScript files. Page look weird? Font look wrong? Something not working right? Post form missing? Try clearing your cache.

 

[rel=alternate]

Not sure if this by design or not, but I only have the option to like profile posts (reaction_id=1)

If I change the reaction ID in the URL, I am still able to choose whatever reaction I want though.

 

[Baraadmirer]

• There is an issue where blank pages are being sent randomly. This also affects fonts, styles, images, and JavaScript files. Page look weird? Font look wrong? Something not working right? Post form missing? Try clearing your cache.

That doesn't answer my question.

 

[Kosher Salt]

I don't know if this helps - probably not - but when the blank page glitch hits when applying a reaction (basically it just sits doing the loading animation for a really long time and then says that an error occurred), the reaction actually does get applied. It'll appear if I reload the page.

 

[Null]

I'm aware. That's why there's so many doubleposts. I have exhaustively looked into it and I have no clue.

 

[WhiteNight]

Not sure if this by design or not, but I only have the option to like profile posts (reaction_id=1)

If I change the reaction ID in the URL, I am still able to choose whatever reaction I want though.

This probably means the JavaScript didn't load and it fell back to being non-interactive. One of the solutions above such as hard refreshing (shift-reload or control-reload) or clearing the cache should fix it.

From CenturyLink thread; moved here since not a CL issue:

I have Xfinity. I have not been able to access KiwiFarms through Private Internet Access -> US Denver for the past several weeks. I actually thought the website was down this whole time.

You should watch Telegram; if the site were down, it would be announced there.

It might help Null a lot if you perform a tracert to kiwifarms.net when connected to the US Denver PIA location. I'd reckon he can't do much troubleshooting this without that.

I can confirm PIA Denver cannot reach the Farms. Traceroute:

 1  _gateway ([redacted])  35.547 ms  35.494 ms  35.458 ms
 2  38.32.95.33 (38.32.95.33)  36.642 ms  36.630 ms  36.593 ms
 3  be2530.ccr21.den01.atlas.cogentco.com (154.54.88.29)  35.936 ms  35.945 ms be2762.ccr22.den01.atlas.cogentco.com (154.54.88.81)  35.926 ms
 4  be2402.rcr21.b006467-6.den01.atlas.cogentco.com (154.54.88.46)  36.407 ms be2401.rcr21.b006467-6.den01.atlas.cogentco.com (154.54.88.42)  36.328 ms  36.321 ms
 5  gtt.den01.atlas.cogentco.com (154.54.10.122)  97.396 ms  97.372 ms  97.342 ms
 6  ae12.cr7-dal3.ip4.gtt.net (213.200.120.106)  54.900 ms  56.287 ms  56.332 ms
 7  * * *
[etc.]

 

[Toolbox]

Most of my other issues have solved themselves, but it seems whenever one's done I get another. Actually managing to quote posts by replying is a bitch and a half, been so for a couple of days. Pages mostly load fine other than an infrequent blank page issue, 429s very rarely hit, but when I click reply the site loads for up to or over a minute and says it can't. Ratings are iffy too.

 

[Aidan]

Using tor browser with clearnet site.
Getting a bunch of 504s today.

Just acknowledging that 504s have reduced substantially for me today.

I usually have a few tabs open and sometimes the need to refresh for Kiwiflare to do its thing results in me being unable to interact until all the tabs are refreshed after one finishes kiwiflaring. Personally not too bothered by it but I can see this causing confusion for people if they leave a session open in the background for a long time then go to interact with a thread.

Throwing onto the cache pile as well but ctrl+shift+r fixes that very easily while the issue persists. Phoneposters need not apply.

ps reminder I'm on the tor browser which may be relevant

 

[Romeo]

Are we getting DDoS-ed?

There will be times where after browsing the site for awhile, the site just loads very slow and it breaks to what @WolfeTone is showing in the image. I don't have a lot of tabs on Kiwi Farms. I then Ctrl+F5 numerous times to clear cache and the site just doesn't load for me, and I feel that KiwiFlare is somewhat stuck after reloading cache from an ignorant point of view.

I'm using Brave Browser Private window and I feel like if it's not coming from the site itself, it's the Brave Browser crapping itself because I noticed how maintenance is becoming worse.

I also sometimes get a 504 error.

 

[Jacknife]

• There is an issue where blank pages are being sent randomly. This also affects fonts, styles, images, and JavaScript files. Page look weird? Font look wrong? Something not working right? Post form missing? Try clearing your cache.

Clearing cache only works for a ver short time, the formatting issue always comes back and stays.

 

[Null]

I don't have any issues and I'm not getting enough information to do anything more. I have thoroughly scanned the back and frontends and I've come up with nothing. I spent an entire day changing config settings in nginx to see if anything helps and it does not. You can stop tagging me about this. Nobody I talk to has any idea, there's no information on StackOverflow about any issues that are even similar.

Does the page look fucked up? CLEAR YOUR CACHE. If it looks fucked up sometime later, DO IT AGAIN.

The 504s mean the backend server is not responding in time. I've ruled out the DDoS guard because it happens without it.

 

[pikachudidnothingwrong]

I'm getting bot check on almost ever new page or refresh again.. (like on day one) so 9 times out of 10 I can't even post!

EDIT: OK never mind, it has once again stopped. Still not sure what set it off but it's happened twice since KF started.. Thankfully only lasting about ~10 minutes each. (usually when just getting on or coming back after a while)

 

[Null]

The site had crashed because I was violently debugging haproxy.

I have implemented a fix. Let me know if the issue with the issue with the cache keeps happening and persists past a cache refresh.

 

[6 Years Ago]

Let me know if the issue with the issue with the cache keeps happening and persists past a cache refresh.

I've been here for an hour this morning posting/reading with no issues (yesterday the site was largely unusable for me even with 20+ cache/cookie clears)

 

[Null]

Okay. There's three caches which complicates things.

File Bucket -> Backends -> Frontends -> Client

The bucket (S3-esque but not on AWS) is how we have multiple backends that share attachments and avatars. It costs money to do transactions to the bucket, so we cache on the backend. Then there is another issue of having to proxy static content to the different frontends (which there will be more of soon) consumes a ton of bandwidth, so the frontend also caches. The client itself doesn't want to ask for the same files over and over again, so it also caches.

If the backend provides a 0-byte response to the frontends, but it includes headers with cache information, the frontend caches that error and distributes to clients who also cache the error. The client may clear their cache but if the frontend is distributing an identical error due to pure shit luck, then it will simply deliver the same error under identical conditions until the cache refreshes.

I've changed it so the frontends will not cache very small files anymore. I believe there might be an issue where the backend sends a zero-byte response from the S3 bucket if it's too slow to download it in time to satisfy the frontends. The 0-byte responses to requests for non-static data (webpages) might be a separate issue caused by the database.

A multi-frontend multi-backend S3 storage cluster like this is an enterprise setup that you usually have a team of very educated and well-paid people monitoring and managing. But lucky you, you have me.

 

[GenociderSyo]

It's definitely working better today I'm editing an op with a bunch of media which was impossible to do over the last few days.

 

[Null]

It's definitely working better today I'm editing an op with a bunch of media which was impossible to do over the last few days.

Were you having the issue where you'd hit the Retard page very frequently?

 

[GenociderSyo]

Were you having the issue where you'd hit the Retard page very frequently?

Only when updating image heavy threads or opening multiple tabs to go through my alerts. Not just randomly like others stated.

 

[pikachudidnothingwrong]

An unexpected database error occurred. Please try again later.