There's a few "RaffleFinished" ones failing, but I doubt that's related.

I would double check the failed jobs list and make sure there’s no notification jobs erroring or failing. Otherwise it should fire assuming you have the config setup. Maybe a screenshot of your notification setup would help.

it's as simple as it gets really

That's with only me in there

If I add the corp itself and another director, they work reliably, but are spammy

Do you see the notification for the extraction under the character itself in Seat? (Character -> Notifications)

yeah, looks fine

yeah, I'm pretty sure that blacklist isn't working... 😄

hm, just updated my containers, and now the horizon page just gives an Error 500, with no further logs anywhere.

Stuff seems to be working fine though other than that, and jobs are still running

https://bpa.st/2H5A https://bpa.st/QOWQ

Is there a complete list of all apps available for the auth?

Apps as in?

This is the most complete list I know: https://eveseat.github.io/docs/community_packages/#package-list

Question, is there a way I can look at exactly when a character was added to our seat, without having access to the cli

click on the User button on the sheet

Yeah, horizon page is still broken, and I doubt it's my fault

I just want to make sure. Our Host Machine is running Debian, so for your Seat5 Manual pull, there were 3 different "versions" of PHP that we could pull. Can we use these same commands to install it on Debian rather than Ubuntu?

docker or bare?

Bare

Ah no idea.. just bloating up the host at that point

There's a debian repo for php that works mostly the same as the Ubuntu one

Can you link just to make sure?

On my phone atm

So no

Its the same site, iirc, just for debian

Finding the corp mining tab is very inaccurate and slow?

Yeah, just went and downgraded step by step. This error first occurs in the seat 5.0.43 docker image. The 5.0.38 image works fine (.39 to .42 don't exist).

# cat /etc/apt/sources.list.d/php.list
deb https://packages.sury.org/php/ bookworm main

fwiw

Does nobody else have a broken horizon dashboard on the latest version?

got a problem i think

nope mines working

Well, it reliably gets broken with above error when I update to >=5.0.43

i dunno which version im on anymore

if you didn't update in the last 9 days, not 5.0.43

2 mins ago?

Then you're on an affected version

Well, it's just a plain error 500 for me

but i do have pulse.. so

re-login as ceo/admin toon?

What would that have to do with an error 500?

It's clearly an issue in the code

if you dont have permission to access said code

that's not how that works.

If there were lacking permissions, you'd get a 403

not a 500, which turns into a massive stacktrace when you enable debugging

ok yeh 403 / forbidden

Or configuration or server status or any number of things

.docs.troubleshooting

@user_301981661761896449, https://eveseat.github.io/docs/troubleshooting/

All errors like that generate logs so instead of just saying it's broken, find out why....

In this case check your laravel logs

 Warlof\Seat\Connector\Exceptions\DriverSettingsException: Parameter guild_id is missing. in /var/www/seat/vendor/warlof/seat-discord-connector/src/Driver/DiscordClient.php:109 

getting dicord error

Im using seat-docker so i dont have a var/www/vendor folder

perhaps this is a database error ?

You do, but the folder is inside the docker container.

You have to configure the guild (server) id

I have. it just suddenlt stopped working one day

Iirc (it's been a while) it should be under the connector settings page

When do you see the error generated?

Was working great for awhile. And today I set it to strict. And updated sets and it started failing every time

Hmm I'll have to have a look later and confirm where it gets the guild_id filled from and get back to you.

Also by failing every time I assume you mean the apply policies jobs

Yes

The error log is pasted above already. It looks to me like a file is missing from the image somehow.

https://discord.com/channels/821361165791133716/821361546608508938/1245075608804003890 this one

I wouldn't know what I configured wrong to cause that. There's nothing particularly special in the config.

And again, everything, including jobs, seems to be working just fine. It's just the horizon status page that's throwing a 500

It's an interesting one. As we don't specifically call out the production of that file. Horizon itself is pulled in as a dependency and configures itself. We don't use vite in the main seat project.

Can you try the command docker compose exec front php artisan optimize

@user_301981661761896449 I got seat up and running. Now just left is that notification table issue

https://puu.sh/K7E6z/6c7bc17438.png

That ran without any error, but it's still the same error.

Inside of the frontend container, the directory looks like this:

seat@c53d90410969:/var/www/seat/public/vendor$ ls
datatables  log-viewer

These are all the manifest.json files in the entire container:

./var/www/seat/vendor/eveseat/web/src/resources/img/favicon/manifest.json
./var/www/seat/vendor/twbs/bootstrap/site/static/docs/4.6/assets/img/favicons/manifest.json
./var/www/seat/public/web/img/favicon/manifest.json

https://github.com/eveseat/seat/commit/cfdfa72226f9682f7cf65bc3a75f5795683797a7 this change looks very suspicious. Adding /public/vendor/horizon to gitignore.

Everything I can find on google suggests that something, somewhere, needs to run "vite build", which would be called by "npm run build", but I can't see either being installed anywhere.

Nah, installing npm and running that doesn't fix it either

I think the main issue is that "php artisan vendor:publish --tag=laravel-assets --ansi --force" does not copy the horizon stuff

https://github.com/eveseat/docs/blob/30c7cb8356a22477ba3012f9a2c17101f32e44a8/docs/upgrading/from_seat_4_0/bare_metal.md?plain=1#L111 here it clearly shows it doing it. However, when I run it, only the log-viewer one is copied

Yeah, in vendor/laravel/horizon/ there simply is no public dir it could copy

https://github.com/laravel/horizon/commit/b86742130940de1fe1eb672b93a8e18d6696fc68 this is the commit that broke it

Not sure if that'd only work with laravel/framework 11, and it needs updated?

seat is still laravel 10

well, laravel 10 clearly does not like the above horizon commit

the issue is that updating to 11 will probably need a major version bump so that plugins can adapt

I'm not disputing that it might be broken, I just haven't looked into it

Then laravel/horizon needs downgraded at least one patch-version

it's pinned to exactly 5.24.4, which removed the manifest it's looking for

pinning it to 5.24.3 instead might be all that's needed for now?

Honestly, I haven't looked into it. But if you have a solution, please submit a PR that fixes it

It's a pure guess, I don't know how to test that in my setup

https://github.com/laravel/horizon/pull/1438 this explains the reasoning behind the change. But I know nothing about the whole framework, so no idea if that inlining is a new feature

Oh hey, I fixed it I think

I cleaned out all files in .../storage/framework

and then restarted the stack

Yeah, that did it

Guess something in there did not get regenerated like it should have been

if i kick my bot from discord. do i have to have every one reconnect there discord through seat or does it save that information ?

hello boys, does anyone know character_notifications table under which schedule? its not update sometime.

Probablly esi:update:notifications

But more likely it has to do with your jobs: Are there jobs related to notification failing and are there jobs related to notification succeeding?

Hi folks, wondering if someone can help me figure something out. I'm trying to install SeAT for the first time, manual install, using nginx / php-fpm. Everything seems fine, until I try to log in using SSO. Then I get a `Client error: POST https://login.eveonline.com/v2/oauth/token` resulted in a 401 Unauthorized response`` error. Logging in with the temporary admin user, the SSO settings tell me that the Client ID, Client Secret, and Callback URL are all 'OK'. Any ideas on what to look at? Thanks

what permissions did you give /opt/seat-docker/custom?

https://discord.com/channels/821361165791133716/821361546608508938/1235705425563750500

The current issue i have is that the Corp Logo will not load for docker..

gotcha. for me, I can't load any of the custom styling

I figured I'd try your fix (to see if I can get some styling to work) because I think I'm having the same permissions issue you had

thing is I don't know what permissions to give that folder

how do you clear the cache in seat-docker

640

no, worker logs not see about notifications request, no succeed or fail, in /horizon/metrics/jobs, i cant find about notifications eveapi route, is this normal?

I can't seem to get the custom styling to work so here's more info if anyone can help me out:

• I did create docker-compose.override.yml in /opt/seat-docker/ and added the text in the Styling guide (and yes I removed the

  's)

• I created custom-layout-mini.css and custom-layout.css in /opt/seat-docker/custom

• I edited custom-layout-mini.css and just copy/pasted the example in the Styling guide (changing only the corp ID in the URL for our corp logo); custom-layout.css is empty (the file exists bit there is no text)

• I ran "docker compose -f docker-compose.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up -d" after I did all that

Still, the login page is the same as the default. Would appreciate any advice. Thanks.

Hi all I have a problem with the SRP plugin. The data is there, but the plugin says "Invalid killmail address" Killmail - https://esi.evetech.net/v1/killmails/118278098/a870f66be49f94ebf1307a4578383ab767f6a8cb/?datasource=tranquility Another killmail, for example https://esi.evetech.net/v1/killmails/116760340/b6ffa2de9a8987841f54694211bf61062e82d73c/?datasource=tranquility everything fine

pretty sure the plugin wants a zkill link, doesnt it?nvm, different plugin

https://github.com/crypta-tech/seat-srp

Well did u run docker-compose.override.yml as well? docker compose -f docker-compose.yml -f docker-compose.override.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up -d

Name the file docker-compose.override.yml

What Matt said above..

But like i said before... The custom-layout-mini.css for the login page to place your Corp logo doesnt work on docker... Whoever wrote the code for the is only looking for the file seat.svg only.. and i havnt found a way to override/overwrite that file in docker... I can only get it to work in bare metal by replacing that seat.svg with my corp logo that was converted from png to svg (it looks really bad that conversion) and called it seat.svg..

QUEUE_BALANCING_MODE = auto 

give me a bad gateway after restarting the stack. Does this not work ?

As well as

 QUEUE_WORKERS

Hello, thank you for getting back to me. Not at first. However, I did do a LOT of combing through this Discord channel last night and I did see someone suggest running that command so I did run that eventually (they also suggested running docker compose down first, which I did as well). Still, that didn't fix it. This -f docker-compose.override.yml command isn't mentioned in the guide https://eveseat.github.io/docs/styling/ so I didn't mention it here in the list of things I tried thinking that it wasn't relevant. Still, I appreciate you trying to help out. Anything else I can try?

Hello, and thanks for the help! I already did this.

One thing I just noticed... perhaps this isn't the default view... it looks slightly different but I can't put my finger on it. I inspected the page and found this:

Perhaps it is trying to load the CSS and it's just not looking how I expect it to because it can't load the images? Sorry, I don't know much when it comes to CSS/HTML. Our website is https://seat.unx-d.com/ if that helps.

And thanks for taking the time to assist me.

in you docker-compose.traefik, there is this line:

• "traefik.http.middlewares.seat-security.headers.contentSecurityPolicy=default-src 'none'; connect-src 'self'; script-src 'self' 'unsafe-inline' 'unsafe-eval'; style-src 'self' 'unsafe-inline' https://fonts.googleapis.com https://fonts.bunny.net https://snoopy.crypta.tech; img-src 'self' data: https://images.evetech.net https://img.shields.io; font-src 'self' https://fonts.gstatic.com https://fonts.bunny.net; manifest-src 'self'"

replace it with this:

• "traefik.http.middlewares.seat-security.headers.contentSecurityPolicy=default-src 'none'; connect-src 'self'; script-src 'self' 'unsafe-inline' 'unsafe-eval'; style-src 'self' 'unsafe-inline' https://fonts.googleapis.com https://fonts.bunny.net https://snoopy.crypta.tech; img-src 'self' data: https://image.eveonline.com https://images.evetech.net https://img.shields.io; font-src 'self' https://fonts.gstatic.com https://fonts.bunny.net; manifest-src 'self'"

The key change is the addidtion of https://image.eveonline.com in the image-src section

That looks promising.... I'll try that now.

And then again docker compose -f docker-compose.yml -f docker-compose.override.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up -d

Let me know if that helps if not I will look later on this after work.

I appreciate that. It seems it did not load any images.... I'm seeing that same error again when I inspect the page. Here's the change I made (I commented out the original line, and added your line just below it):

Here's the part with the relevant change

OK I will check this out later

Thanks again for the help!

Hey I got the background to work!! I Added https://web.ccpgamescdn.com to your line. Now it seems I'm just having the same problem as @user_255905434970357761 where the corp logo doesn't load. Matt thank you so much for steering me in the right direction!

For anyone that might be searching this channel for the fix after following the guide: Open /opt/seat-docker/docker-compose.traefik.yaml with a text editor like nano. Comment out the line that starts with traefik.http.middlewares.seat-security.headers.contentSecurityPolicy=default-src 'none'; connect-src.... with a

(see my screenshot above)

Just below that line you commented out, add this entire line: traefik.http.middlewares.seat-security.headers.contentSecurityPolicy=default-src 'none'; connect-src 'self'; script-src 'self' 'unsafe-inline' 'unsafe-eval'; style-src 'self' 'unsafe-inline' https://fonts.googleapis.com/ https://fonts.bunny.net/ https://snoopy.crypta.tech/; img-src 'self' data: https://image.eveonline.com/ https://web.ccpgamescdn.com/ https://images.evetech.net/ https://img.shields.io/; font-src 'self' https://fonts.gstatic.com/ https://fonts.bunny.net/; manifest-src 'self Once you save that yaml file, run these two commands in /opt/seat-docker/ in this order: docker compose down docker compose -f docker-compose.yml -f docker-compose.override.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up -d That'll at least make the background load. It seems getting the corp logo to load is a bit more difficult.

Yeah I was going to modify this to add this website to the line.

Seat logo is bricked so yours is displaying but underneath of the seat logo 😛

I have to check how I did my fix

Hey, the background changing is huge, at least... looks super cool so I'm pretty happy as it is.

so adding

QUEUE_BALANCING_MODE = auto

or

QUEUE_WORKERS

seems to break my seat installation. After removing them from my env file still downt work and i had to reinstall seat, any idea 's ?

@user_175284736095158273 can u show me your css?

for logo

if I'm seeing this correctly its from https://images.evetech.net/

correct

1 sec

/* SeAT login page layout */ @media all { html, body { height: auto; } .login-page, .register.body { color: rgb(255,255,255); background-image: url(https://web.ccpgamescdn.com/aws/eveonline/sso/background.jpg); background-position: center center; background-repeat: no-repeat; background-size: cover; background-attachment: fixed; } .login-box, .register-box { width: 360px; margin: 0; position: absolute; top: 50%; left: 50%; background: rgba(48,48,48,.8); transform: translate(-50%, -50%); border: 5px solid

ecf0f1;

border-radius: 40px; box-shadow: 0 1px 1px rgba(0,0,0,0.05); } .login-logo, .register-logo { font-size: 35px; text-align: center; margin-bottom: 25px; font-weight: 300; margin-top: 50px; } .login-logo::before, .register-logo::before { content: " "; display: block; width: 128px; height:128px; margin: 0 auto; background-image: url(https://images.evetech.net/corporations/2040769686/logo?size=128); border-radius: 50%; margin-bottom: 50px; } .login-box-body, .register-box-body { background: transparent; padding: 20px; border-top: 0; color: inherit; } }

it's basically exactly the same as the example in https://eveseat.github.io/docs/styling/ except I changed the corp ID in background-image: url(https://images.evetech.net/corporations/2040769686/logo?size=128);

I will pm u with a fix as .login-logo is not defined in new seat.

i'm interested in that too for a bare metal install

Sure I will send it to u later when I get home

is there a easy way to take a docker database backup and use it for a bare metal install ?

Sharing is caring 🙃

shouldnt we be able to just add the image into

curl -L https://raw.githubusercontent.com/eveseat/seat-docker/master/.env.example -o .env Gives me a 404 not found in the .env file

nvm. Found it by hand

using which instructions?

Its during the upgrade process

When downloading the template

I guess the example file is not with traefik in mind right?

I'm interested in that too

hey man it worked for me and it wasn't all that hard! mind if I make a guide for it here using the steps you told me?

hi, i'm trying to pull our own copy of seat5 for manual install for edits. i'm getting a whole bunch of "composer file says php:^7.2 but you have php 8.2" are there any ways to get around this so I dont have to edit every composer file? this also spans outside of php there are other reqs like larval -framework and such. How does seat5's composer get past this?

are you sure you have the code for seat 5? Seat 5 only supports php 8

since seat 4 uses php 7, I suspect you still have the code from seat 4 installed

i pulled Seat-web along with a few other projects.

@user_164235653008064513 I'm 99% sure you just pulled the master branches of all the projects, which aren't the latest version

yeah master is still 4.x

Forking only pulls master

you can fork the others into your fork

Also "Forking only pulls master" is only true on GitHub if you leave the checkbox labeled "Copy the master branch only" checked

a fuk yeah, ok thats most likely the problem ty ty

I dealt with forked project, but I've never actually initiated a forked project myself. ty ty

https://discord.com/channels/821361165791133716/1247605206372323359

@user_189293982952456192 for bare metal only the directory will change I believe

changing the url to a file on my host for the logo doesnt work in bare metal. But it did work for the background

Did u and an entry to a volumes?

Docker-compose.override.yml in my version has an entry to add for logo file

The logo is getting overridden somehwere in a bare metal install

it is working on bare metal

wget that in ../public and follow the instructions from @user_405130542674083842

Is there anythign special that I need to do if the seat plugins are not showing up?

I added them after the fact and they simply are not showing up

I took down the stack, added the plugins to the .env file and then brought the stack back up

Did u add them correctly?

0/ The job_batches database has grown greatly. Can I truncate her safely?

you should be relatively fine truncating it, but first you should ask why did it grow this large?

do you have a lot of failed jobs?

look at your failed jobs in horizon

I am sitting stable at 390k failed jobs per 7 days now...

Yes, I periodically clear the queues. 30-40 thousand accumulate in a few days.

Almost everything PlanetDetail

Thank you! Cleared job_batches Everything seems to be working.

@user_614098468218339348 Guys, second question. I have 2000+ characters in my database. I am gradually coming to the conclusion that there is no need to keep some of the characters in the database. I would like to reduce the load from unnecessary requests. I would like to mass delete based on certain criteria: 1) The character is in the NPC corporation. 2) The character is attached to a base that has been in my corporation for more than a year. I wouldn't want to do this manually. What's the best way to do this? Perhaps it's worth developing a plugin? Who would take it?

Right now, I don't think there is a user-friendly way for this

Hi guys, i have a small issue with manual install of seat5 when we connect to the website by IP it works perfectly fine, but when we try to connect to it with the cloudflaredomain it doesn't seem to work. the settings in the .env and settings for nginx all point to the cloudflare domain. any idea what the issue is?

wait i'm dumb LOL chrome auto completes https and i dont have the certs set up

NVM

what gives you the impression it doesn't work? any errors?

it works when i remove the https with http

Just remember.. if you are using cloudflare proxied.. it's effectively useless if someone can just direct connect to the host in question, If cloudflare ranges are the only ip's able to connect then it's secure behind cloudflare proxied.

Hey just a quick question here. I just set up SEAT 5 and without having permissions assigned to specific characters they are able to see all characters authed to the SEAT instance. Interesting here is, they have director roles in game while characters without arent able to see all characters. Is this expected? Cheers!

Did u set yourself as admin?

Myself yes - but the characters arent linked to my admin main

On your main admin go to specific character that u want to check. Press user button and check what permissions they have.

No admin

well it doesnt have to be an admin permission to see all

so check all permissions that you gave to those chars

Ah sorry for. Let me check.

Ok, got it fixed

what is the command or how do i remove packages from my bare metal SeAT 5.x install ?

is it just put seat into maintence mode and remove it from my /var/www/seat/vendor folder ?

If they are directors they are allowed to see all character data related to their own corp..

Or CEO, just to make the above clear

Mhm!

yep yep - found out the other way 😄 Thanks for clearification!

will more cores help on seat? we have 6 core atm and its close to topping out

how many tokens?

3370+

that's respectable, what's your

seat:buckets:update

set to?

30 mins

and it's maxing 6 cores?

what's your hosting arrangement?

AMD EPYC 7282 16-Core Processor, 6 cores

VPS or VM on a dedicated host?

vps

yeah that thing is oversubscribed to fuck

lol

so.. vps issue?

I run 4200+ tokens on 2 cores and 10GB ram with 20 minute bucket update

or we pushing it to hard

ok

Ryzen 3700X CPU

so it sounds a lot like a VPS issue

have got 28 workers atm

how's this looking?

(I have a shitton of failed citadel jobs for god knows what reason...)

i did do an update about an hour ago

right

and i dont think its hit a refresh yet

at */30 you should have hit at least one refresh

ive dropped the workers down to see if the cpu load drops

if it dont, then its the vpn

how much memory do you have?

16gb max used is 3

though we are using 91gb hd space

but i cannot figure out what the db size is

heh

i dropped it by 10 workers

... help i dunno what happened...

ok i did a package update, was still working, then server reboot

yeh i dunno

ok seems my traefik file corrupt?

phue... i gotta stop poking stuff

just noticed this in the startup

is that this ? or elsewhere

i just added them in there* they didnt exsist

narf it didnt fix it

| Cannot create cache directory /home/seat/.cache/composer/repo/https---repo.packagist.org/, or directory is not writable. Proceeding without cache. See also cache-read-only config if your filesystem is read-only.

so ive tried what i could find on it, and none of the chown methods have worked on it. im thinking that cache issue is causing some of my issues

----

fml i hate this shit... i cmod'd the acme folder..

hello you heroes of the eve world - any known possible update probabilities on compatibility for the buyback module for seat 5?

#channel_1113334797201317941 is epic, thats awesome, is that a plugin / bot / native feature of discord?

I have been actively trying to get in touch with the dev about that

no response thus far

some exceptionally talented people here been doing this for years and making miracles happen - i have zero doubt that one of em could do all of give it a sidelong glance in effort and it magically at least works heh

frankly if I don't hear from the proper dev, I'm totally game to put a bounty on updating it to v5

I have a time pressing need for a buyback plugin

and I am more than happy to pay someone for their time to do it

I hope you are using docker engine and not docker desktop too.

It's a native discord thing but you have to convert to a community server to use it. https://support.discord.com/hc/en-us/articles/360047132851-Enabling-Your-Community-Server

My buyback fork is working : https://packagist.org/packages/hermesdj/seat-buyback

I think i'm having issues with the notifications -> webhooks -> discord flow

i've setup two notification groups, one for killmails and another with 'all' category and hooked it to two separate chats, however nothing is coming through on discord so far

you effn super hero - @user_143815222740647937

THANK YOU

😄

There is also a new configuration to allow a "the corp buy EVERYTHING at -X%"

And it uses the new price providers

you guys should update this as the goto community package for this

any working one over a phone that wont pickup is the winner imo

omg wild. thank you so much

10/10

is there a way I can force a full update on a target character? when new characters get added to the system there seems to be a significant delay between them being added and their information being available to peruse... which is slowing down recruitment due to needing all the info for BG checks.

do I need to setup a discord bot or will a webhook and a webhook URL matched to the notification suffice?

Tests go through fine but nothing else does.

I found the esi:update:characters character_id command and have successfully run it. the batch shows up in laravel, but it completed instantly and there is still no data. which makes me assume that the tasks were skipped due perhaps a caching setting or something?

Might be caching or maybe scopes for the tokens of this character ? No info not public info either ?

yea nothing

and I have characters that have been in the system for over a day and are missing sections of data like contracts

from what I understand, the refresh button doesn't actually spin off API calls and instead just refreshes the view from the database?

like I could really use the ability to tell the system "hey, despite any internal management systems, make the API calls and update this data"

scopes for the characters are the defaults that come with the installation, which I believe should include everything needed

If you need a full update the command you used is the only one I know of. On some pages there are refresh buttons but they only refresh part of the data afaik

If you look at the batch do you see any errors ? Maybe look in the logs of the worker ?

yea no errors in laravel. I haven't dived into the console logs too deep yet, but I didn't notice any errors pop up while experimenting earlier

it would be nice to know what endpoints the various tasks are using so I could try to do some raw experimentation with ESI to confirm I'm not crazy or something

Well to know the endpoint you could look at the code directly : https://github.com/eveseat/eveapi/tree/5.0.13

Like this one for example : https://github.com/eveseat/eveapi/blob/5.0.13/src/Jobs/Character/Info.php

I'll do some digging then

I'd look into the logs first, there might be a big error there

kk. yea cause out of like ~60 characters there are only 2 contracts listed in the system

so either something is broken about the contracts stuff specifically or idk

Need some help tracking down why I've got an overheated process

215% cpu from systemd-coredump through mariabd - but there's nothing under /var/lib/systemd/coredump. Restarting both vps, and the docker container sees the CPU drop, for a few minutes, but spikes back up around the 5 minute mark. Also not seeing any errors in syslog, and process list is empty.

delete them from the system and have them relog in - this is how i deal with the issue toons when the waves hit

that's just not feasible with how often this is happening

I understand the desire to make sure the application is respectful of ESI, but I need the administrative ability to tell the system to run the API calls

hi. i install seat 5 on ubuntu server 20 and when i run 'su -c 'php artisan seat:admin:diagnose' -s /bin/sh www-data' all looks goods. i can loggin succesfully but don't syncronize any data when see the characters. any idea why this happen?

Question: what would be the best approach to enable external access to the mariadb? I want to access the db in MySQL Workbench, so connection to the server over SSH, then locally it'll need to connect to the Mysql server. I assume disable skip-networking and then binding to 127.0.0.1 in the my.cnf, that should be safe enough right?

are you on docker or bare?

docker

just upgraded our install to 5

anyway to stop the system trying to grab the same stuff and failing over and over?

its failing because its tried too many times...

did you figure this out? I might be having the same issue. it seems that about every ~7.5 hours or so the cpu usage skyrockets to the point where I can't even reach the site

I think I fixed my issue. It was running it on a droplet with 2gb of ram, which didn't seem to be enough. rigged up a swap file and it seems to be doing way better

I have not

anyone knows how to repull/overwrite or install a fresh container of seat-docker-front-1?

My notification web hook is sending same killmail twice

is there a way to increase the page size when making an API call? or is it locked to 15?

It is set by CCP and varies per endpoint

What affililation do you have set up. If the affiliation matches multiple times it will send multiple messages

Can you be clearer about your intent? donw then up will recreate the container

It's same affiliation as you can see same killmail is posted

I would have to check at the code.. But I think its becase both the victim and the ceo/director of the corp can pull in killmails, so its coming from both

Are you referring to the high attempts count?

Easiest path, is make sure your server firewall blocks port 3306, then add a ports declaration to the mariadb compose file to expose the port. Then mysqlworkbench can login over ssh and connect locally

are your supervisor jobs up and running / working correctly?

The command here is mariadb not coredump.. Its understandable to be a heavy cpu load on the core mariadb processes itslef as SeAT is really nothing more than a data collector and manipulator in the DB.

i dump the server and make a fresh install from scratch. i will install seat4. thx anyway.

As a reminder seat 4 is no longer under development

oh. damm. ok. i wll install seat5 and if have a problem i will post here.

Out of curiosity, 6 core processors being maxed - is it just polling too much? Table pulls look all to be sub ms so quite quick but maybe indexing issues?

yeh its still trying them over and over 2500 ish of them in 36h

I've also been seeing similar behavior since the changes to fix the updated_at values that were previously causing assets to appear/disappear. I'm sure it's related somehow, but I haven't had a chance to look into what's causing it. The screenshot here is all a single character ID - this one has quite a few assets. All the other tokens seem to process fine (~250) - just this one fails with seemingly varying numbers of "attempts". Rarely, it will succeed, which is why I haven't dug into it super far yet, but the job does seem to run for the entire hour (+5 minutes or so) before failing. Guessing it has something to do with the batch for the character running hourly and the job not completing prior. (8vcpu/16GB RAM)

every one of these is a different char, some are at 47, other 60's and a few 70's and servers been up 38h now

sorry if this has been asked here before (couldn't find any info on it but wasn't entirely sure what to search): is there a way to change the description that shows up in the third party apps page in the Eve website? just want to change it to something more obvious like "SeAT Corpname"

The owner of the application should be able to update it on https://developers.eveonline.com/

ah, too obvious! thanks man!!

Corps are not updating correctly. I try to add a schedule for it and get this error. Has anyone seen this before?

restart your containers and all jobs you need should get added in case you removed one accidentally

Still no esi update corporations update.

hi again. im rying to install mariadb (this is under seat 5) and get this error. any idea of what is going on? i search for a solution on internet but any solution i read works. thx in advance

so then let me ask about the wallet transactions one specifically. in the UI it's possible to increase the page size up to 100. can I do that in the seat API as well?

That command does not belong in the scheduler like that

Hmm so when someone joins a corp the auto squads do not always add them even though it shows them in the correct corp/alliance., It used to work but stopped recently.

What are u using as your filter?

Match kind: All Type: Alliance Operator: is Criteria: "Alliance Name"

Classified? and Auto on the squad settings.

Deleted all auto filters and re-added. This seems to have fixed it.

Not sure if this was fixed yet or not but just for reference I think it’s known. https://discord.com/channels/821361165791133716/821361546608508938/1228300478643703869

Not related, but also fixed

Looks like apt get doesnt have the latest version. I can only find 1:10.3.39 there https://packages.ubuntu.com/focal-updates/mariadb-server Maybe try following the steps here for ubuntu 20.04 https://mariadb.org/download/

Hey all, I just updated to seat5 and I don't see errors on the docker logs. Nor do I see any failed jobs in the horizon dashboard. However, the new system logs shows a large amount of esi-related errors. Primarily 403 errors. Now I am just wondering if the 403 is a result of authorization issues, or if its just the esi setting a boundary. Any thoughts? Thanks in advance

403 means forbidden error, so it seems the tokens are invalid

If you are not seeing failed jobs in horizon, I assume that the system-logs you refer to are the eseye logs? If so could you show the example logs. My suspicion is more on citadel jobs

my citadel errors stabilized around 396000 per 7 days 😄

steady at that

probably not ideal

Yeah correct eseye and laravel logs are giving the errors. I have dm'd you a snapshot for eseye and laravel respectively

There is nothing overly damning in those logs other than ESI being ESI and timeing out a fair bit. The 403s are to citadels are also unfortunately normal

Is there any one who can send me a log of a failed Citadel job log that is not the max attempts exceeded error? @user_131158012831203328 ?

Do you see anything in the laravel logs coming that may be coming from these jobs?

I'll have a look, but work is kicking my ass hard right now

lot of this,

and char batch processing

Access denied to https://esi.evetech.net/v2/corporations/635210548/assets/locations/?datasource=tranquility {"exception":"[object] (Seat\\Eseye\\Exceptions\\EsiScopeAccessDeniedException(code: 0): Access denied to https://esi.evetech.net/v2/corporations/635210548/assets/locations/?datasource=tranquility at /var/www/seat/vendor/eveseat/eseye/src/Eseye.php:322) [stacktrace]

ah found the asset one

huh i updated last night before bed, after the sde is listed as updated

and yet its saying its not

Thanks very much for looking into it! You have put my mind at ease 🙂

New has been released

Update again I guess

yeh , been finding workers add exponential increase to db load. ive had to cut workers down to 16 to get it below 90% cpu load

thx for the answer. i try this and finally works install mariadb. any other problem i will post here. thx so much for help. 👍

hi guys, me again. finally finish install new seat 5 from zero and have this problem. when i try load the page have error 500. run diagnostic command and get this. any idea?? thx in advance

is there an easy command to requeue an update for every call for a specific characterid?

basically I find (occastionally) a character be missingn like...their security status or skill points

implants / attributes

have no clue how to re-queue those jobs

should say...when hitting the api on v4, to the v2/character/sheet/ endpoint I get a local.ERROR: Trying to get property 'balance' of non-object {"exception":"[object] (ErrorException(code: 0): Trying to get property 'balance' of non-object at /var/www/seat/vendor/eveseat/api/src/Http/Resources/CharacterSheetResource.php:77) which i attribute to some jobs having not finished, or esi barfed on them

i went through an manually did a bunch of the esi:job:dispatch, but I donmt think I got all the right bits as it didnt fix the issue

I also had this issue and wasn't able to find a way to guarantee that the ESI calls were actually being made. But, my problem has seemed to have sorted itself out significantly after I realized that my SeAT instance was maxing the system out on RAM. The droplet I'm running it on has 2GB of ram. after I added a 2GB swap file it seems to be running solid without issues, and the missing character info seems to have resolved itself

yea, just trying to figure out how to recover from a brown oput

I've been seeing this problem for about a month now. There are a lot of characters whose tokens are not updated. The request to link the character again does not yield results. Is there a solution to this problem

are your jobs being processed?

yes, and now I see errors only in assets and citadels

so i thought it was a minor miracle that someone had a 5.0 buyback extension working, and thank you sooo much for that.... so now i ask, ... anyone have the corp skills / certs plugin working on seat 5.0?

solved by deleting the current token on the CCP website and repeated linked character

Hmm, weird

seems my server getting same problem

hi threre! I successfully updated seat now. But after that i see that. Does this module require a separate update procedure? How to do it?

generically php artisan eve:update:sde

thx much

question, I want to run an instance of (bare) install of SeAT5 while running (docker) SeAT4 on one hosting server. The goal it to keep both up for a slow transition. reason why we don't want to migrate the database is to "start a new one" Issue : Both bare and docker argue about port using port 80 and 443. I tried modifying bare's sites-available/seat to point to a different port other than 80. And I also tried modifying trafik to point to another port other than 80/443 like 81/8080. Is this how I am suppose to do it or should I be modifying the docker application port itself. If I am doing this correctly, than is it assuming that its cloudflare that isn't directing the ports properly.

you're going to have everyone re-reg?

you'd prolly be best figuring out how to map the docker ports to alternate ports, and then have your bare install use the normal ports. problem is that normal http/https traffic isn't going to work for alternate ports. and if you're using something like cloudflare... then yea you'd have to see if it's possible to somehow configure port usage there as well. not to mention if this is hosted on a rented server, you need to make sure the alternate ports are available to begin with

imo you'd be much better off standing up the second seat instance on an entirely different machine

unless of course you actually own the machine and have 100% full control over how it is setup and the internet and whatnot

I still fail to see the benefit in ditching the old database and tokens

you're adding a lot of friction for users

I tried to separate several services running in Docker, I came to the conclusion that it was easier to split them into different virtual machines 🙂

our existing database is sort of in a "corrupted" state. every now and then we get the ajax error when people try to search for assets. a couple of threads were made about this and no real solution was presented, so we might as well just start new.

we also get those from time to time, I attribute it to weirdness in the UI/UX code

do you still get them in SeAT5?

I haven't tried

I have other issues in SeAT5, concerning citadel jobs, and without looking at the code, I suspect it's some sort of weird-ass regression

Any help is welcome.. I can't see why the jobs are failing the way they are tbh. If the citadel cannot be resolved, then the job should just complete normally and not fail the job.

I mean I expect a lot of failed esi calls... But not failed horizon jobs now that I am debugging deeper.

If someone can provide me with logs from failing citadel jobs that are not the Max attempt error then that would help

Multiple services within docker is very simple. Especially with traefik as a reverse proxy.

Moving ports is fine, so long as you also point your browser at the new ports.

everything I see in horizon is "attempted too many times" ...

but my install has been running steady-state for a couple of weeks now

Horizon is probably not the best way to look at it as you will only see a fraction of the jobs. Like looking for a needle in a haystack. May be easier to grep the laravel logs. Or use a plugin like Seat-pulse to view aggregate stats

well, sure

but how does the logic look like for these retries?

like it cannot be intended behaviour for jobs to fail in this manner

"too many retries" ... what triggers the retry? and on what schedule?

That's what I have come to realise, and am hence trying to debug. Originally I had thought that the jobs were failing because the citadel could not be resolved. As that is how I initially laid out my original proposal of the job. But that's not the case here. According to the job code as I read it, we should catch the exception and if it's the invalid access exception just succeed the job anyway

I'm getting a lot of 503s and 504s ... of course, because the ESI is shit

And for a 503 or 504 I would expect the job to fail

Then be retried

oh, I just noticed this one, not seen that before...

[previous exception] [object] (PDOException(code: 22007): SQLSTATE[22007]: Invalid datetime format: 1292 Incorrect datetime value: '2016-03-27 02:29:00' for column `seat`.`corpo
ration_medals`.`created_at` at row 1 at /var/www/seat/vendor/laravel/framework/src/Illuminate/Database/Connection.php:587)

😄

Interesting. Will add that to the later thoughts list

(ie I will forget about it and deal with it later if I remember or a gh issue reminds me)

  less eseye-2024-06-13.log |grep 403 |grep structures |wc -l
12551
cat eseye-2024-06-??.log |grep 403 |grep structures |wc -l
82215

yes I'm a bad...

but that's a lot

heh

And that's expected in the eseye log. We want to see that we are getting the right responses from ESI in the case you don't have access

yeah

Is there anything in the laravel log from the Citadel class?

[2024-06-13 03:18:16] local.ERROR: Seat\Eveapi\Jobs\Universe\Structures\Citadel has been attempted too many times. {"exception":"[object] (Illuminate\\Queue\\MaxAttemptsExceededException(code: 0): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel has been attempted too many times. at /var/www/seat/vendor/laravel/framework/src/Illuminate/Queue/MaxAttemptsExceededException.php:24)
---
cat laravel-2024-06-13.log |grep -i citadel |wc -l
53674

OK, can you exclude the MaxAttempts and see how many lines left?

let's see now

typical output:

#1 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#2 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()
#5 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#6 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()
#1 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#2 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()
#5 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#6 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()
#1 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
--
cat laravel-2024-06-13.log |grep -i citadel |grep -vi maxattempts|wc -l
248

Hmm... Sorry to be a pain, but can you swap the order of the greps and add a preceding 2 lines to the citadel grep?

yeah ofc

[2024-06-13 06:44:05] local.ERROR: The datasource tranquility is temporarily unavailable {"exception":"[object] (Seat\\Eveapi\\Exception\\UnavailableEveServersException(code: 50
3): The datasource tranquility is temporarily unavailable at /var/www/seat/vendor/eveseat/eveapi/src/Jobs/EsiBase.php:484)
[stacktrace]
#0 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/EsiBase.php(317): Seat\\Eveapi\\Jobs\\EsiBase->handleEsiFailedCall()
#1 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#2 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()
---
cat laravel-2024-06-13.log |grep -vi maxattempts |grep -B3 Citadel|less

lots of 503 and 504, mostly unavailable

[2024-06-13 11:45:58] local.ERROR: The datasource tranquility is temporarily unavailable {"exception":"[object] (Seat\\Eveapi\\Exception\\UnavailableEveServersException(code: 503): The datasource tranquility is temporarily unavailable at /var/www/seat/vendor/eveseat/eveapi/src/Jobs/EsiBase.php:484)
[stacktrace]
#0 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/EsiBase.php(317): Seat\\Eveapi\\Jobs\\EsiBase->handleEsiFailedCall()
#1 /var/www/seat/vendor/eveseat/eveapi/src/Jobs/Universe/Structures/Citadel.php(97): Seat\\Eveapi\\Jobs\\EsiBase->retrieve()
#2 /var/www/seat/vendor/laravel/framework/src/Illuminate/Container/BoundMethod.php(36): Seat\\Eveapi\\Jobs\\Universe\\Structures\\Citadel->handle()

Hmm...

Do the numbers of 403 in eseye logs line up with the laravel logs?

seems to be a multiplier

laravel has 4x the errors of eseye

Hmm what about for 503 and 504?

let's have a look

503: 100 504: 131

As in that's the total number?

that's the total number of those, from june 13

403: 12551

Roger... So we are definitely getting the 403... I just don't get why we are seeing it retry...

yeah, iirc it's supposed to blacklist those, right?

Less that but just the job retry is indicating the error code is not reporting.. Which means something about the way the job completes is not correct

Which may mean it's not reliably hitting the blacklist also as a byproduct. But that's more a symptom than a cause in this case if that makes sense

yeah

hope all the logs and stuff was helpful to your debugging, I need to get some stuff done now 🙂

Yeah thanks heaps. It helps me to work through where things are going wrong, as it's very hard for me to replicate

sudo -H -u www-data bash -c 'composer update h4zz4rddev/seat-buyback'

Does not update package.

Worked for the SDE but doesnt seem to update plugins

are you pinned to a version?

unsure what you mean by that ?

Unfortunately mine are mostly “too many times” as expected so probably not much help here, I do occasionally get 503s though with the same error as Wibla but mine gets about ~2000 total and then clears up once the previous jobs fall off in a week as the cache is established. My personal curiosity has been related to how the resolution works when you have 50+ citadels (or even just multiple) on the same batch and one or more fail with a 500 - just a thought but possibly completely irrelevant.

just to confirm what do you mean by this

do you mean the incoming browser ie chrome pointing to the new port?

Does the low fuel structures also alerts pos towers?

or just the upwell structures?

Right - as in you’d bind one or the other to a different port (at least temporarily) and point your browser or reverse proxy at the specific port instead.

No, just upwells. I think it was removed long ago in anticipation of the removal of POS towers but as we know those were never removed.

https://github.com/eveseat/notifications/blob/0b70be0c19a39ba61c46b1162c8d06e36e9141f2/src/Notifications/Starbases/Mail/StarbaseFuel.php#L33 its still there, kinda. Not sure if it can be easily reactivated

@user_270988688807297026 how do i add for that for docker?

🤷‍♂️

Ok @user_270988688807297026 I assuming you can install this on a Bare Metal? if so how?

> Not sure if it can be easily reactivated

After 24 hours these tokens turned yellow again 😦

What scopes are associated with the tokens?

I did not understand the question

Are the esi scopes for the token set to only publicData?

no, full Only alts turn yellow tokens; mains (users) do not have this problem. those. only those characters that are added through this button