wonder if they didnt have a price source imported from v4

Ahh! Sorry yes that would do it. V4 didn't have price sources. Apologies I forgot about that. So after an import from v4 to v5 need to specify price source.

sall good

Growing pains

Hehe. I don't think about v4 a lot anymore

Starting more to think about a potential v6

remind me to write out a wish list

Throw a curve and just say it's V7

ok teamspeak, settle down

I'm getting some wild runtime numbers with some character related jobs that report back that the software has hit the error limit for ESI. my task queue has been struggling to keep up over the last couple days because of it. any advice?

can you send us your eseye logs to see which requests to esi fail?

Since that last update 3-4 days ago, were getting near 200k errors a day again, before it we were down to 2k a week

Are you talking about the eve update or a seat update you installed?

In general, esi has been quite bad in the last few days from what I hear from others

killmail esi endpoint looks like boom

Well at least that one will get fixed fast by ccp because of zkill

lol

A lot of ESI endpoints have been quite disruptive over the last 3-4 days, mostly timeouts. Seems to be majority killmails but also is affecting other jobs like orders, status, history, etc. Feels like it's been roughly since the ESI outage last week but I haven't looked at the timeline close enough.

U right, I stop market endpoint, and now needed to stop killmail

are there steps on where to find the eseye logs somewhere?

and regarding turning off endpoints, does that involve just deleting the related scheduled task?

.docs.troubleshooting

@user_614098468218339348, https://eveseat.github.io/docs/troubleshooting/

also on the web ui you can access the logs browser in the settings drawer. settings->log viewer

gotcha

yea the eseye logs are pretty much completely 504 gateway timeout errors

I can export one of them for you if you need it

the types of calls failing are all over the place

order history, killmails, skillqueue, contracts, planets

there are a bunch of 404s for a calendar route https://esi.evetech.net/v3/characters/483957378/calendar/401753/?datasource=tranquility

there's the one for today if it helps

looks like it's been having these errors since the 17th-ish

yes, that is basically your only option

if you need to readd them, just restart the docker stack

kk

Whats the line for adding workers to the server for SeAT 5?

I copies over the old one from my SeAT 4 deployment, that's not working 😅

It's the same

QUEUE_WORKERS=20 in the env file

Do I need to put it in a certain spot int he .env file or naw? Because I have this already: QUEUE_WORKERS=40 that's working as intended, apparently. but uh....laravel is only ever showing 21 max now total processes even with maxed jobs and I have over 180K failed jobs and batches taking sitting in pending for more than 10 minutes. Just have over 187K failed jobs and lots of batches are shown as sitting in pending past 10 minutes, which matches my previous SeAT 4.x schedule and workers.

Just trying to understand the 5.x and optimize 😅

So.. 40 is a lot of workers..

But keep in mind esi is very broken at the moment so everyone's jobs will be failing and now is not a good time to optimise things.

Ah okay, thank you. I was hoping it wasn't a "yeah you done fucked up your config*

I can prolly tune that down to 30 on 5.x, it seems to run jobs A LOT faster than 4.x did lol

do we have a statement from ccp? or do we know whats up?

I guess they ran out of money to pay for their shitty cloud environment

😛

used to sit around 300k

lol

I'll be real honest... I think you are going to hit a resource constraint before you hit a worker constraint and adding more workers just slows all of them down

I rarely went over 6-7 GBs of RAM on my 4.x build with that many. Server is reasonably beefy.

Hey, unfortunately it's your disk that is in question probably

okay that's nuts. How many cores? how much memory? good storage?

Also how many tokens

that too yeah

also what's your bucket update set to?

400gb nvme SSD, 8 cores, 16gb of DDR4 RAM. VPS. I think we're at 2200+ characters? 10 minutes bucket updates. 40 workers.

Used about 41gb over the last 3+ years.

so it's a VPS

shared infra

That's the resources dedicated to my instance but yes.

"dedicated" 🙂

those are always oversubscribed to a certain extent

Allocated*

fair

there will be memory pressure, cpu contention, storage contention etc

It's not a US or UK provider, German one, Contabo. We got it back when their stuff was super cheap. Prices have gone way up for new customers since sadly.

It's in their US data center though.

I would be onboard with that if I consistently had the same error count with the same settings on 4.x but I didn't 🫠

I had my buckets set to 8 mins with 4.x, actually.

I only ever had to clear the eseye cache when ESI itself shit the bed fantastically.

how long have you been on v5 now?

....actually now that I think about that I don't know how to do that on 5.x, I'm not sure the same container command works?

lol

❤️ docker ❤️

Less than 4 days.

right

that's after CCP ESI shat the bed, iirc

so you don't really know heh

Ah okay I didn't know that happened in the last few days, just lined up right with my upgrade. Wonderful

When Crypta said earlier "ESI is being bad" I was like "....right when I updated??" but....this is CCP we're talking about. Sorry for prolly coming across as doubting you, Crypta, if you read this.

I held off on updating for so long because I knew how to work 4.x well enough, but some of the plugins I was using stopped working, figured it was time to update.

I think it's healthy to have a suspicion of updating to a new version, tbf 😅

I'm running keydb instead of redis, eseye cache in redis, 18GB ram, 4 vCPUs on a 5600G, 2x Samsung PM683a enterprise SSD in ZFS mirror on proxmox, 4400ish tokens and 20 minute refresh

lots of errors the last few days, but it runs pretty well

that host basically only does SeAT and a couple of other things though

Yeah mine is overkill for just SeAT

But it's only $16.99/mo

🤔

All the US/UK providers were charging double that in 2021

I mean you've backed off the bucket update intervall

By two minutes 😂

I hadn't realized there was a long-going (relatively) ESI issue

well the default is 2 minutes

which, considering the sheer write amplification that happens with SeAT, is pure insanity 😄

Yeah lol

Worked fine for <500 tokens

Gotta get back to work though

Yep 🙂

hey! going through the api documentation, i am trying to figure out how to request a different page from a request. I get the meta data "currentpage" and "lastpage", and am struggling to find a way to move to page x

looking at /api/v2/users currently if that helps

you can use query string like this: /api/v2/users?page=12

amazing thank you!

Well issue with shared even if you are guaranteed X amount of ram it doesn't mean much on shared hosts. Also you say 16GB of ram.. but saying you're using 41GB just means sooner or later if that's your ram usage you will just be banned or removed as that would be considered excessive on a shared host. Then again some shared hosts could give zero fucks if you over utilize.

I meant 41GB of SSD storage, my bad

How do you clear eseye cache on SeAT 5?

Or can I do that through the scheduler now

I used to do it via command line lol

It hasn't changed

dope

Okay, sorry, tried running that old command you told about in DMs over a year ago, just modified for the new docker compose: docker compose exec seat-web php artisan se:ca:cl Got this

root@vmi618539:/opt/seat-docker# docker compose exec seat-web php artisan se:ca:cl
WARN[0000] /opt/seat-docker/docker-compose.override.yml: the attribute `version` is obsolete, it will be ignored, please remove it to avoid potential confusion
service "seat-web" has neither an image nor a build context specified: invalid compose project

wait seat-web isn't even a container in 5.x I'm dumb. Justr trying to clear the eseye cache

The container names have changed. It's now called front

ty

Not sure what this means when trying to clear the cache?

root@vmi618539:/opt/seat-docker# docker compose exec seat-docker-front-1 php artisan se:ca:cl
WARN[0000] /opt/seat-docker/docker-compose.override.yml: the attribute `version` is obsolete, it will be ignored, please remove it to avoid potential confusion
service "seat-web" has neither an image nor a build context specified: invalid compose project

okay renamed the override file and not I get this:

root@vmi618539:/opt/seat-docker# docker compose exec seat-docker-front-1 php artisan se:ca:cl
service "seat-docker-front-1" is not running

But....

[+] Running 6/0
 ✔ Container seat-docker-mariadb-1    Running                                                                                                                                   0.0s
 ✔ Container seat-docker-cache-1      Running                                                                                                                                   0.0s
 ✔ Container seat-docker-front-1      Running                                                                                                                                   0.0s
 ✔ Container seat-docker-worker-1     Running                                                                                                                                   0.0s
 ✔ Container seat-docker-scheduler-1  Running                                                                                                                                   0.0s
 ✔ Container seat-docker-traefik-1    Running                                                                                                                                   0.0s

that looks suspiciously like a mix of seat 4 and 5. seat-web doesn't exist anymore in 5, it is now called front

also you can ignore the warning about the version being obsolete

you exec by service name, not container name, so docker compose exec front php artisan se:ca:cl

AHHHH okay

Awesome, that worked, ty

the old command was seat-web so I imagine the service was called seat-web

Aha

I understand better now

use containers they said... it'll be so much easier they said... shit I upgraded and now my commands don't work anymore, halp?

Use bare metal they said… it’ll be so much easier they said. Shit I upgraded and now I have to roll back to a machine image backup, see you tomorrow. 🤣

snapshots exist

:v

or just do a PBS backup before updating v0v

I guess if you're on some shitty VPS that's problematic

is ESI broken? got a huge amount of timeouts..

Has been for two weeks on and off now

Not just ESI.. clients are also dropping

yeah

had all my clients drop offline before DT

good times

Well that's lucky.

I can't stay connected for more then 20 minutes or less.

with multiple titans undocked?

but yeah, haven't logged in much the last week or so, been busy with work

Or trying to dock in a keep or npc and disconnecting each time you attempt to dock

Sitting under pos shields right now

damn

that seems pretty fucked up

I'm effectively coffined right now

CCPleeeeeaaasse

it got even worse :v

just looking at this

yep, esi stability is not at it's best that's for sure

2.5 weeks

tbf, EVE stability isn't all that great either

getting all my clients disconnected at the same time at irregular intervals

We've folks from the UK area that gets dcd at the same time on daily basis. I'm wondering if the global problem is related to network issues or if it is their pure infra instability

Iv heard recently Verizon and Cloudflare had network issues...

Mostly likely just Verizon fuckery.

lol

that is some shit

WHAT CCP DOING NOW

https://tenor.com/view/the-simpsons-homer-simpson-spider-pig-pig-ceiling-gif-4466277

i thought mine was looking bad

Unfortunately there's not a lot we can do to improve the quality of esi.. The avenue we have been advised by CCP to take is to bring up the issues we are having with the CSM.

any idea why seat slows to a crawl when we have this many errors?

Most likely ram pressure. Unfortunately laravel decides to store failed jobs initially in redis, before a flush to the db after 24 (iirc). The downside of this is lots of ram pressure which can then itself cause more errors. (positive feedback cycle is bad)

A method to check that is to run php artisan se:ca:cl to clear the redis cache out and alleviate the ram pressure temporarily

ram seems pretty stable at 80%

you mean docker compose exec front php artisan se:ca:cl

ok that took its sweet ass time, but ram dropped a lot

"You all don't need api right?"

since doing the big update last night, error's have dropped down to 5k in a day

oh yeah?

tha'ts interesting

Yeah mine is showing just over 200k failed jobs 😕

800k ...

https://tenor.com/view/it-just-works-tod-todd-howard-%D1%82%D0%BE%D0%B4%D0%B4%D0%B3%D0%BE%D0%B2%D0%B0%D1%80%D0%B4-fallout-skyrim-gif-14803660

esi is slightly better atm. you can try clearing your job queues and cache to see if things inmprove

I've noticed my errors have slowly decreased through the week from 200k to almost less than 100k

mine were ok.. till overnight

fml ccp

looks like things have improved a bit at least

i did a update clear an hour ago

Question, what if we can help their load by decreasing the request by 10%, 25% or 50% (im just pull numbers randomly)? or if CCP can come up with a new esi recommended standard/protocol? just saying if we call contribute, it could help them?

maybe they had to scale back their cloud infra costs 😛

so now it's just overloaded and shit in general?

i am so annoyed, SeAT is not even posting some killmails twice....no...some KM get not posted at all!

we slowed down our refresh from 2 mins to 30 mins and it still goes

in 3h

1000 in 5 mins

and 2 mins

i guess esi is dead

Sound like you going to chip in to CCP cloud infrastructure 😉🙃

I have 14 accounts subbed

I'm doing my part.

givenn your esi usage though....dont think you break even >:)

p sure I do

😛

not my fault that CCP chose a literally braindamaged architecture for their API

then claiming it was the only way to decouple it from the monolith

I call BS

and as long as your DB is happy, nothing prevents you from adding more workers, even on a second node

i'm not so happy; it takes 120 hours to update everything once

im still trying to find right balance

the new squads update though has helped a toooon

recursive really nocked it out of the park

BTW: did the last update resolve the oom issue?

yea

no ooms even after I dropped it back to 128meg

if the Esi would stop puking all day every day maybe we could get some normal error numbers again

holy shit

what kind of bucket update schedule do you run?

oh I see it

120 hours

ooof

Buckets are seat 4+

i modified it to a rolling schedule - the next batch of char queues when the previous batch finishes

seat3 😮

so it's not really scheduled in a sense - the scheduler runs per minute

what kind of hardware does it run on?

epyc 7601p 64 core with 32G ram allocated on web

and 128G ram allocated on database

but the db is not dedicated to this. there are tens more sites on that machine as well

~37k 504s in the esey log yesterday....down from 65k the day before

gettin better every day!

dayum

you dont have prune running?

v3 dont have prune

?

thats not so useful

it only deletes orphan stuff

supposed to prune dead tokens

fml

those are valid tokens

there are another 40k dead tokens

¯\(ツ)/¯

nope. i use the bootstap script. which leads to the access denied error. if i attache to the mariadb container and change the permissions for the seat user. the frontend container is showing this error.

i am not sure if this a follow up error because of the access denied error or if this is an error on its own.

okay after a little bit of tinkering. 1. clone the seat-docker files manually 2. set all values in .env 3. boot mariadb container 4. change the permissions to seat@'%' 5. boot any other container but front 6. boot front this is working for me. i think there is an issue with the mariadb user permission setup procedure. i am running on a debian latest, docker latest, with "docker compose"

did you have seat installed before?

https://eveseat.github.io/docs/installation/docker_installation/#manual-deployment

nope. clean new VM

you should not need to do anything like that on docker

true. but this was necessary

and it is not working if the front is started once before changing the user permissions. only if i change the permissions before i start anything else.

first installation try was. 1. boot a clean new debian 2. install docker https://docs.docker.com/engine/install/debian/ 3. bash <(curl -fsSL https://git.io/get-seat) -> no success second installaiton try: 1. boot a new clean debian 2. install docker, git 3. clone repo 4. ./bootstrap.sh (without acme because i do a opnsense sll termination endpoint for ssl) -> no success third: 1. boot a new clean debian 2. install docker, git 3. clone repo 4. up -d mariadb 5. change user permission seat@'%' 6. up -d -> working

it's very very veryyy strange, looks like you've already a volume attached to the mariadb container. also you don't need to create a new vm brute force mode between all the install just be sure to down the seat stack via compose and remove the volumes (on a fresh install right) don't do that if you've a previous install or you'll lose all the datas.

so step by step via the bootstrap script you shouldn't have to clone anything. As the script will make eveything you need during the .env setup process which is takes ESI infos, instance URI, mode (proxy/traefik) and that's it the script will generate all you need to start

Access denied for user 'seat'@'172.18.0.4' (using pa  
front-1      |   ssword: YES) (Connection: mysql, SQL: select table_name as `name`, (data_le  
front-1      |   ngth + index_length) as `size`, table_comment as `comment`, engine as `engi  
front-1      |   ne`, table_collation as `collation` from information_schema.tables where ta  
front-1      |   ble_schema = 'seat' and table_type in ('BASE TABLE', 'SYSTEM VERSIONED') or  
front-1      |   der by table_name)

this is the message i get if i down and up again or do a clean first boot

you need to delete the docker volume to get a fresh DB

tried that.

even with docker system prune --volumes

and do a complete overhaule of all images and networks...

no change

Does docker system prune delete docker compose volumes? I'm not 100% sure. Can you run it and then check if there are any volumes?

no volumes

i tried to step down to 5.0.50. booted like it should. changed back to tag 5. now it works. absolutly no idea why.

The easiest way to remove the volumes will be to run docker compose -f docker-compose.yml -f docker-compose.mariadb.yml down -v

then up the stack again

Hi. Am a SeAT noob in the sense of setting up and running an instance for our Corp. Installed via the bootstrap script. Everything loaded and worked first time. Now I want to expand the instance with some plugins. Trying to follow the plugins instructions and failing at the very first hurdle - cd'ing into /opt/seat-docker. I keep being told that there is no such directory. As such, I cannot get to the .env file to add the plugin name. Not too sure what I am missing after re-reading the installation documentation. Any advice would be appreciated.

Try using ls /opt/

See if the seat_docker folder is there..

Can't remember if you need to use "sudo" before the "bash".

docker container inspect front find the compose label area, It'll say where the container's compose is in question.

Thanks for the replies. ls /opt/ returns - digitalocean docker container inspect front returns - Error response from daemon: No such container: front Going through my file system with Filezilla, I have found /var/www/myauth, so attempted to run the Blade Package Installation instructions, being in the supposed install path, with: sudo -H -u www-data bash -c 'php artisan down' This returns bash: php: command not found Installed php-cli............ Now I get Could not open input file: artisan I am kind of guessing at present it may be simpler to reinstall SeAT from scratch..........

Having /var/www/myauth present makes me think you have allianceauth installed. Are you sure you installed seat?

What gives ls /var/www/myauth?

> Having /var/www/myauth present makes me think you have allianceauth installed. Are you sure you installed seat? Yep - been running it for over a month 🙂 Ass aid, frontend works fine. I just can't get into the backend to do anything. Attempted to install Alliance Auth originally, but it failed. The VPS droplet was completely reset and SeAT installed. > What gives ls /var/www/myauth? ls /var/www/myauth returns static

Output of find / -path docker-compose.yml?

Does not give me a return. Goes back to prompt. I am going to do a reinstall of the instance after resetting my droplet as I can not even get back into Admin, which I have been into after initial setup to set access on who can see the ESI data, etc. Might be a case of something DigitalOcean has done (I remember they did a PHP update not long ago) that might have thrown something out? Thanks for the assistance and I might be back later 😁

Hm...i'm getting an error for "Container seat-docker-traefik-1" while they all are spooling up. 🤔 "Error response from daemon: driver failed programming external connectivity on endpoint seat-docker-traefik-1 (b4135fcaef6a10f238e6a78eb2c2492bc141fe361c145de6e9dcab2c63ba88cc): failed to bind port 0.0.0.0:80/tcp: Error starting userland proxy: listen tcp4 0.0.0.0:80: bind: address already in use"

do you have any other web server running on that machine?

tbh, its probably two instance of traefik running

I'd do a docker ps and see if theres a dangling container from a previous run

If not, ``sudo ss -tuln | grep ':80'`` ``sudo netstat -tuln | grep ':80'``

as long as the Batch is not installing one of its one...nope

Scheduler-1 / Woker-1 / front-1 / cache-1 / mariadb-1 shows up. Does not look like a dangling container. all have the same start time

then its astrals command will tell you whats listening on the same port

root@ubuntu:~

sudo ss -tuln | grep ':80'

tcp LISTEN 0 511 :80 : root@ubuntu:~

sudo netstat -tuln | grep ':80'

tcp6 0 0 :::80 ::: LISTEN Well not really getting wise from that. Just that its listening on tcp6 but the container is looking for tcp4?

do netstat -tulpen | grep ':80' instead

tcp6 0 0 :::80 :::* LISTEN 0 13052 741/apache2 The Apache2 comes up.

there ya go

hm...next question would be for a noob like myself...why and how do i stop it 🤔

would have to know why you were running it to start with

is there some other service running on that server?

Nah...but according to the Documentation, there is a WebServer needed. NginX or Apache or not? And if i got it right, PHP: ^8.3 want Apache2 or not?

everything is inside docker

nothing external is needed

ah so the fun things down is there for the manuall shit

yea

if nothing else needs it, I'd just uninstall apache

well...i had a feeling about that already. but my first try with only the docker was also a miss, maybe i did something stupid there 😄

containers are the way

Depends on the host os.. some have apache installed and running by default

Hmhm...its running now. Is it right, that the Admin Link is always a direct link that put me into Seat without the chance to "create" a specific account with specific credentials?

Yes

Sign in with your main

then sign in with admin link

give your main admin

Profit

hmhm...how long does it normally take to fech the prices for stuff like the moon reporter?

o/ seems our seat is bloated, could do with some help 'unbloating' it if possible

woah, how many tokens is that, and how long timeframe?

20 month, 1600 char in corp

not sure on token count

it seems like it is mostly the DB. Can you give us a breakdown of how much space each tables uses? Also, 50GB doesn't seem that crazy. I know the requirements in https://eveseat.github.io/docs/installation/requirements/#up-to-1-000-characters are probably very old and maybe no longer accurate, but 1.6*20GB=32GB vs 50GB doesn't look very crazy, given it also grows with time

got a cmd line to do that? i havent had much luck trying to see that

curl https://www.google.com/search?q=mariadb+get+size+of+tables

1. connect to your DB (different on docker and bare) 2. run this sql query: https://mariadb.com/kb/en/information-schema-tables-table/#view-tables-in-order-of-size

this is all i ever get back

and the backup is only ever 2-3 mb

i get nothing but syntax errors trying to use the commands from that site

copy and paste it, syntax error, type it in lowercase, kind of works

Do the structure attack notifications work for metenox moon drills? I see that notifications are sent to the characters

Looks to work

Our tokens wont refresh since 1 Day. Complete SEAt of the corp. Are there Problems sometimes on CCP side or died our connection? Diagnose says all good: * Checking ESI Access Server Online Since: 2024-10-17T11:02:06Z Online Players: 24462 ESI appears to be OK Checking Local and Latest Versions. Please wait... +--------------------+---------------+----------------+ | Package Name | Local Version | Latest Version | +--------------------+---------------+----------------+ | SeAT API | 4.9.0 | 5.0.0 | | SeAT Console | 4.8.0 | 4.8.0 | | SeAT Eve API | 4.21.0 | 5.0.16 | | SeAT Notifications | 4.3.3 | 5.0.9 | | SeAT Services | 4.3.1 | 5.0.8 | | SeAT Web | 4.20.0 | 5.0.18 | +--------------------+---------------+----------------+

Seems susspicios. Or what means inactive.... the Jobs on the second page are running?!

is the only Error i can find. somebody can explain or tell me why seat isnt updating the Tokens?

Simple. Because horizon is inactive. It is what orchestrates jobs to be processed

I'm assuming this is a manual install, if so check your supervisor config is running

https://eveseat.github.io/docs/installation/manual_installation/#supervisor

ok will check

seems true. will have to check why....

I've sometimes had to kick that thing into gear twice

i restarted it 3 times now..... but still failure

maybe sde outdated or so....

i dont get it running with systemctl restart supervisor.service I didnt changed something in the last 3 months...

My seat server not working when clicking menu "SeAT API"

It redirect to "http://mydomain.xxx/api-admin" but it shows 404 not found

You set the email right?

It worked for months..... today it seems to recover without any hands on. i dont know what happend.

2nd entry

Yeah I know, So if your stuff is setup on seat it should now show the seat page.

we will see in a bit

that ended up bricking the domain lol

me linking the domain to seat now causes the website to no longer load

I am at a loss like how does me linking seat cause the website to no longer load

Your instance of SeAT is not served by "your domain". It is served by "seat.yourdomain" as configured by your DNS A Record and the Ngix/Apache config file.

I know when I tried https://seat.blackwatertech.space/ it shows that traefik is there but maybe misconfigs somewhere.

well I just deleted the entire docker again

the folder

Careful.. you can only request TLS certs every so often

and the website still has not recovered back to the default hostinger one

Are you installing through bootstrap script?

yes

I am using the bootstrap script on the vps

to install seat to the vps then I am trying to connect to it on my actual machine

okay it reverted back to hostinger

after deleting the docker folder again

but soon as I point seat to this domain using the bootstrap script the website no longer loads

neither does seat.blackwatertech.space

it just does that dnsprobefinished_NXDOMAIN

Dude

Did you set up that subdomain?

no the steps I did was buy the vps, buy the domain, then run the bootstrap script

on the vps

and also set up the eve developer application

like I said I have minimum website experience so I am looking for info what to do after running the bootstrap script but I cannot find any info

Then it won't work

"it doesn't work, I have tried nothing and searched for nothing, please fix my problem"

hey like I said I was not aware on what to do after which is why I am asking here

You need to create a subdomain

right

This is DNS 101, find a howto

until you said that I had no idea I needed a subdomain

https://tenor.com/view/alec-baldwin-alcohol-pour-drink-pouring-alcohol-gif-16944581

It didn't cross your mind to search for that error you got?

Don't forget that a DNS update (i.e. your A record) can take some time to propagate on the network. Took ours 24 hours to work. Sometimes it can be quicker. You can check your propagation status here: https://www.whatsmydns.net/. Looking for lots of green ticks 🙂 If red crosses, your SeAT instance will not see the site request yet.

all green apart from in japan and mexico and seattle

no as I thought it just meant the domain was not working

Assumptions

okay so I need a subdomain

Google it

Good luck.

I'm going to bed

wohoo now I need to buy website hosting to go on top of vps hosting and domain

oh I am looking it up now

okay so now I am searching it I assume I need a website as well as a vps and a domain?

No. Sending you a dm kosmos

right

all done

Hmm what's going on there..

they helped me set it up

turns out it was a simple problem with how hostinger handles ip adresses

They guided you through getting the vps and such set up?

the domain and the vps that seat was hosted on used two different ip adresses

changed the domain ip a rule to the VPS IP adress and it fixed it

now just got to go through seat and try to figure out how to use a api checker like what eve hr had or look for a community add on

Sorry what?

hostinger used two different ip adresses instead of one I used a single one and it worked

also got a new problem, to do with installing community packages on dockr 5.0

docker compose not docker-compose

ah

I was just copying and pasting the command

another problem but I will solve it tommorrow is that SeAT does not send any notifications to my gmail account that I listed for the username

works for discord though

Did you set up the SMTP server info?

I did

Just make sure if you are using gmail to send the email you use an app password and not an account password.

that was it

I was using my gmail account password

did not work, it doesn't matter I can still get discord notifications and that is fine gmail was optional anyway but seat is set up finally

From where? The docs should all use docker compose and not docker-compose

@user_614098468218339348 > From where? The docs should all use docker compose and not docker-compose On SeAT's main website. Stated under community package instructions: > Once you have entered the package(s) to be installed run the below command in /opt/seat-docker: > > Docker (SeAT 5.x - using Traefik) > > docker-compose -f docker-compose.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up -d

Ok thanks. I’ll fix it

https://github.com/eveseat/docs/pull/177

for compose it will always depends on docker / compose version depending how hold the server is (up to date)

i guess we'll always find some people having issues with it, while it's just common sens to know which version you have / can use

While you can run seat 5 with docker-compose, the installation instructions and docs all use docker compose

Or at least should, as we’ve just seen

is there a community addon for buyback which is avaIlable for Seat 5.0 ?

the only one on the website is not available for 5.0 yet

The version posted in SeAT Community Packages (h4zz4rddev/seat-buyback) does install on V5 - we have it running on our instance. Not had any issues yet, not to say we might not run into some as not specifically for V5.

okay thanks

found one for 5.0

only caveat is it has no discord notifications

https://eveseat.github.io/docs/upgrading/from_seat_4_0/docker/ I installed docker on windows, is there a windows version of the upgrade guide ?

This part of the upgrade guide is broken

404: not found

I suppose we should update it : https://raw.githubusercontent.com/eveseat/seat-docker/master/.env.example

SeAT on windows is not officially supported. There are no installations guides, so there are also no upgrade guides. If you managed to install it on windows, we generally expect you to be able to upgrade using the linux instructions

I will see how it goes thank you 🙂

thanks for the report. it will get fixed: https://github.com/eveseat/docs/pull/178

it worked well for me it seems, will see if I encounter other problems

only "mv" command, had to be changed to "move", also I think "curl" command is not in standard windows installation but I have it myself.

``docker compose down && docker compose pull && docker compose up -d``

@user_247176974164819968 I want to connect to my mariadb instance from my host. I added a port section into docker-composer.mariadb.yml (that is what I did at your advice in seat 4 and worked well) : ports:

• 127.0.0.1:3306:3306

It does not allow me to connect, there is something else I need to do ?

remove the 127.0.0.1 part if windows.

I assume this is past advice for awhile back?

yes 🙂

I'd say 127.0.0.1:3306:3306 if you were sshing in and doing an ssh tunnel, But if this a windows host on your network just do 3306:3306

docker run on my computer on windows, and I want to connect from my host with a mysql client (heidiSQL for my part) into my mariadb instance

my seat is just for my use to retrieve information about my characters like assets etc

just do what I said above, 3306:3306

127.0.0.1:3306:3306 is only really for vps and doing ssh tunnels.. as some vps providers have the os set to just have any port that is open be accessible to the world the 127.0.0.1:3306:3306 prevents external but allows ssh tunnels to operate just fine.

does not seems to work after restarting the instances. Seat working fine but I have no access from host

You down'd and up'd?

yes still the same thing, but the only thing I did is to add ports:

• 3306:3306

i did not run any ssh command into the instance

What's the connection string you are using?

it's a desktop sql client, configuration did not change

127.0.0.1 is invalid put the host ip

everything is running on my computer

Maybe the firewall is blocking the port?

that wouldn't surprise me

firewall de-activated it is the same thing

lsof -i -n

does not return anything when run inside mariabd, this should return something if a tunnel is created right ?

https://superuser.com/questions/248389/list-open-ssh-tunnels inspired from here 😅

On traefik instance the output is like this

ok I solved it

my docker-composer.mariadb.yml missed the line I highlighted

I guess seat-internal is a private network not accessible from the host system

for some reason, i cant get seat to properly ping a discord channel when a pos is attacked/reinforced. here's my notification group, the discord integration "test" works perfectly fine. had a pos reinforced today but no notification. is the pos reinforced notification an alert type i missed? i didnt see any others that would be relevant in the list

there is no notifications for POS events.

Is there a reason? Just not implemented (if so I can look at doing a pull request) or is there a technical limitation you're aware of? I know there is no pos reinforce alert but there is an attack alert in-game. TowerAlertMsg notification_id

POS's are very old tech. So maybe that's why ?Perhaps they are not impletmented in the ESI anymore

It's come up a few times, I think it was removed/deprecated a while back when CCP stated they were going to remove POS from the game entirely, and either it was removed or just not updated and then later got fully removed since the code wasn't up-to-date and I don't recall it being functional. There are definitely notifications for them still that get captured, just someone would have to put code together to get it working.

I heard it was one guy that did all the POS code, and he either quit or got fired. and ccp couldnt figure out his code so they made the stations we have now instead

POS are still in the game

they are still in ESI

they are still actively used, shot, pinged and memed

I heard the same thing Kiba. If i remember right, the reason was the CCP was suppose to remove the POS from the EVE long time ago when CCP implement Upwell structures. SeAT "removed" the ESI code in preparation for that. So officially POS are "still planed" to be removed. It may or may not happen any time soon or later... Unless CCP announce they will keep it, and who knows if that will ever happen...

For affiliation remove corp and use a director or a ceo

Event will be tracked through you ceo director or any other role that can recieve the notifications for a starbase being attacked

Keep in mind if you add (multiple toons for each corp) you'll have N*(toons) the same notif for a given corp

Yes, pos notifications got removed from seat, but I think they are still in esi. If someone can capture me a few POS notifications, I can readd them

in other tools, multiple tokens per corp means tokens get rotated to get the delay time down...

What kind of notifications? How would you capture it?

I don’t have many but I could grab some from my DB later tonight if that would be helpful or maybe can trigger some new ones lol

look like my clones does not update in seat, is this a know issue ?

I see no error in the jobs

what is the job name for the clones esi update so that I can investigate ?

Literally called 'Clones'

If you monitor tag clone you should see them

Why would the command to do a database dump not work?

just instantly completes

no output

what command are you doing

There a reason for pos events, i was telling this for months, there are corps with more than 40 in space, mainly whormhome corps, the ESI its still sending those alerts, there is no reason indeed in null sec where there are no tower, they are old but very usufull and still in use, and my 10 cents they can't just remove it from the game untill fast anchoring structures are a thing. Seat 4.0 does the notifications, sest 5.0 still have the code just disabled/deprecated. POS's are still a thing mainly in wh space, having the notifications its a most, i have to run one separete app to give me those and that makes no sense.

There are there, they are implemented

Seems true

Exactly, they are a pain to set up but very usufull in multiple situations, even in null people warp caps to it when in combat timers

/corporations/{corporationid}/starbases/ /corporations/{corporationid}/starbases/{starbase_id}/