Sounds like it

Im getting the same issue with duplicates discord messages/notifications.. not sure if it is related to the failed jobs that re-attemp/retry it again...

every time I log in this warning show up, any way to fix it?

In settings page, add an admin email address

top box in there 🙂

ok, thanks, what is this email used for?

Am I supposed to receive some notification from SeAT?

Requests to the EVE API need to have an administrative contact email set before SeAT will queue jobs to process. CCP made the request in a Github issue. To address this, the email address is added to the User-Agent string that is used when making EVE API requests as can be seen eveseat/eveapi:Helpers/PhealSetup

L77.

it's a CCP requirement

ok, I see, thank you

how do I get seat to update my corporation structures?

because seat hasn't updated anything for at least a month

do you have a ceo or director authed?

I have a director authed

Preferably something you will check..

If CCP can't reach out to you for api issues they'll just ban your api in question, which will invalidate all tokens if they do.

No it does not

IP Addresses and Tokens have zero interrelation

depends on what you are fucking up

if you are just blasting the rate limits you'll get ip banned

but if you are doing something weird thats not harmful, but they dont like you'll get an email

No, you wont

CCP discard the email address before it reaches any logging or reporting. But they still ask us to include it, so we do, to be good users.

haha yeah ...

that is hella stupid

"Watch me, put local@local.local"

You should be a good citizen of ESI, then the onus is on CCP.

Well, Being a good citizen also means using a valid email if that's a want from them and not telling people there's no risk at all for hitting the api heavily and restlessly. If there's no conquences who's to say not to use a local address that can't be reached at all and to ignore their rate limits..?

I agree, you are the one promoting a fake email XD

and providing incorrect advice about the relation between IP bans and tokens

Telling people there is no consequences promotes a "Set it to whatever" mindset alongside with "If there's no consequences I can do whatever I want."

Thats a weird out of context ping

but i agree, without all that gymnastics required

I really need a solution to this stupid double posting of notifications...i changed to Affiliation test wise towards the CEO Char and no Notification got pulled at all. As soon as i put it back to Corporation, it sends too many Notifications again. -.- It can not have anything to do with the amount of Directors in the Corp and that he sends Notifications for every Director. Because those Numbers also not match up.

.

@user_214209749900722189, available bot commands: .ping, .docs, .esi, .sde, .version

you've to understand that if you have the whole corp on seat and you've set the notification settings to the CORP, every single toon that will receive that notification will fire a notification event. so depending on what you want to track you just have to set one toon only that recieve the majority of the notification, CEO or Director for example is a good bet and remove the corp affiliation.

IF that is the case were seat get those information from the CEO and/or director, then i think there should be a check so it can only send one notification. For example, Notification check should set primary from CEO, if CEO token doesn't work for what ever reason you can come up with, it should move to the director as a secondary source...

well everybody is free to do a PR with features improvement if you feel like something is missing right 🙂

https://tenor.com/view/suggestions-discord-paper-shredder-printer-gif-21816765

Okay...i was all the time having a brain fuck...obvi the toons also need to be in SeAT yes...BUT...how do i request Killmails then for example? No one gets Kill Mails on a personal level, except the person who killed or lost something.

Separate groups for different notifications

Only CEO for corp things

Corp scope for the personal?

well if i use Corp Affiliation on the Killmails i get the Notification twice... So i would switch all Corp Notifications to the CEO but still...what to use for the Killmails then...

Think the corp also gets them maybe?

Try it 🙂

Any full director has access to the kills for whole corp even those that are not posted on zkill. So yes the kills work like the person who killed gets it and person who got killed got it + directors from those pilot corporations. This is the access in game to the kill.

In notification affiliations u should put only one character preferably with the highest access in game that is registered with your seat instance. Like CEO or full permission director

Yea..just wasn't aware corps got km notifications

Just to make sure we are all on the same page, we need to use the CEO or Director in the "available character"?

Yes. Leave the available corporation empty.

Thanks

Seat keep trying to pull a kill notification from a few days ago and it keeps failing... Also, one of my corp mate just killed a ship with a group and i still havnt seen any notification for it..... why???

If they weren't final blow they don't get the kill in their history, Also you won't see anything like zkill unless they explicitly post it or consent to auto post.

You'll only get the ESI object in SeAT if you have the character token that got the final blow, or a corp key for someone with access to the kills (which still requires someone in that corp to have gotten the final blow), to be more specific. While a Zkill link is generated in SeAT, it doesn't have any direct link to it. Zkill attaches everyone on the killmail to a particular kill entry and has many many tokens associated with it to pull all of these from the ESI. Without a token that specifically gets your killmail, you won't see it in SeAT and thus won't get a notification. However, regarding the above error specifically - it seems like there could be some sort of problem getting populating the ID number for this code here: https://github.com/eveseat/notifications/blob/5.0.x/src/Notifications/Characters/Discord/Killmail.php#L95 I'd guess there's something in the killmail that is unanticipated, like a NPC or an ID that SeAT isn't able to resolve, hence the null error.

Thanks for the explanation.

If you want raw Zkill and not to be limited by SeAT's killmail knowledge, I'd suggest a separate bot that provides filtered output from Zkill directly. A few of these are floating around.

what do you mean by seat.failed jobs? The DB table? They reside in redis for the first day

Yeah I also cleaned the db table after the docker update as it had 200.000 rows in it, but atm there are 390 rows in it with mainly Seat\\Eveapi\\Jobs\\Character... Seat\Eseye\Exceptions\RequestFailedException: Character has been deleted! errors, but those are from 45 minutes ago

I'll wait a bit and try to login into the web front again later, the kid is waking up soon as well... 😉

Lol now I tried in Firefox instead of Chrome and it logged me in

hey, just a quick question: is it possible to get all assets reside in a specific location. So we want to see if someone who did not got the messages have stuff in "old" structures

I have a query to check on jumpclones with implants, undoubtly its possible for assets as well Only downside is afaik it isnt possible with the regular webinterface SELECT ci.name, cjc.implants FROM seat.characterinfos ci INNER JOIN seat.characterjumpclones cjc ON (cjc.characterid = ci.characterid) WHERE cjc.locationid = 1044075195228 AND cjc.implants <> '[]';

@user_164132248554176512 if all your failing jobs are ppublic jobs then I think I have found the issue.. New docker build will be available in about an hour

Thanks! Will check later in the evening

Lol.. i havnt gotten any kills yet.. nor did the Moon Mining extraction completed yet.. or attacks on my structures...

Yeha dunno why it should Work with the CEO. I mean yes...the CEO gets all the notifications also but not the Killmails.

cheers

it figure something out

I'm actually thinking about using another Bot atm...makes just no sense the problems with SeAT here. dunno why its so hard to just send one Notification on Corp Afiliation and why the supposed "just the CEO" is not working for me

Let me know which one you pick.

Maybe https://github.com/SvenBrnn/hazardous-killbot?

It's quite good.

so, 400k failed jobs in 2 days, np right =0? troubleshooting a seat instance with someone

this is fine meme insert here

2400 tokens, 2vcpu, this is fine

my v4 seat has been strugglin a lot tbh

ive grown a lot of tokens recently though

but am about to migrate up to 5 and onto a new server

that could be fine - if the bucket update schedule is setup right

Default atm =0

2 min lol

I do have to say though, with some of the latest updates here to batch squad updates and such from Recursive/Crypta my CPU and disk utilization has gone significantly down from where it has been for many months. I've spot checked a number of items and everything seems to be working as expected. I've never had to adjust my bucket timings yet even in the past, but I went from ~350 tokens regularly using a good chunk of 4-6vCPU to about 2vCPU or less average - and this is with default token scheduling and no changes. Obviously I'm not suggesting to go bleeding edge with releases if you have major important instances unless you are ambitious - but it looks very positive.

I'll have to pass all credit here to recursivetree! They have done an amazing job with these changes (and with putting up with me in reviews / pre-code discussions)

Well, either way, I appreciate everyone that works on it for spending your time on it. Maybe this image will be interesting for recursive_tree to see the impact - blue line is where I applied the updates, just in case it somehow isn't completely obvious.

That Proxmox right?

Reminds me of Grafana

It also looks like the hetzner dashboard

To be honest, I was surprised myself how much better it runs. And someone else contacted me with concerns that the update broke his seat instance because the cpu usage dropped by about 80%

Honestly this was also my first thought was that something was broken - mostly just because I’d gotten used to SeAT being relatively heavy. So far that I can tell though, seems to be running as expected. I’d actually migrated over to a more powerful instance about 3 months ago due to a couple issues, one was that more tokens had been obtained over time increasing load, second being heavy overprovisioning of the original VPS host causing random poor performance and job failure, and third was that EU instances (vs US hosted, even East coast) just seem to not have as many problems with ESI calls in general due to the much lower latency. Now I’m considering cutting the instance in half.

more money in your pocket is always a good thing..

I need to upgrade as soon as I get back online

It sounds like

Yes I'm very curious to see the impact on your system xD

Can confirm my CPU and disk can breathe after this update

Maybe I can scale down a bit more 😎

I'm running into an issue with the Corp Mining Tax plugin to where completed contracts are being displayed in the console as NEW and unpaid. I have already verified the contact description, etc. when the contact was sent were correct.

what seat version?

4.x (latest)

Separately - Is there a way to see which of our corp buildings are generating the incoming from people using them? ie. mfg or BPO research?

@user_164235653008064513 do the needful please, see if that helps us.

no

longer answer: there are tricks you can do with changing the tax that allow you to narrow it down by hand, but seat does not help you

if you are really keen you can look at the index for the system and combine it with your tax logs to figure out no only where, but what they did

I have upgraded now 🥳

probably won't see any huge changes in the graphs before a few hours have passed

Haven't noticed big changes in load

After 13h

Slightly lower cpu scaling

still seeing a lot of writes, I will have to profile it a bit, might be redis eseye bgsave being a dick

CPU load appears to be down (by a bit)

To be fair the performance changes here are mainly related to squads and observer processing so are very circumstancial. It will vary a lot server to server based on character activity and squad configuration

ah yeah, I don't use a lot of that

The main thing that will probably help you the most is the ability for longer and varied schedulues for token pulls

that is interesting, yeah

is there a writeup on that?

Kinda: https://github.com/eveseat/web/pull/684

I should update docs to reflect at some point

yeah... that'd be great 😄

features are worthless if they are not documented well enough that people can actually utilise them

🙂

True... Noting anyone can contribute to docs.. Its one of the lower priority things on my list atm

some thoughts on granular ESI scheduling is to look at what kind of job types takes the most resouces, and run them more seldomly

but I am coming from "4000+ active tokens, already running 20 minute bucket refresh"

Potential.. But the approach here is to filter based on character affiliate groups as opposed to job type

but if I could keep the 20 minute bucket refresh for slow jobs and run less "costly" jobs more often, that'd be pretty interesting

yeah I see that

True

I guess we're approaching this from different angles 🙂

still pondering how to deal with redis TTL in a better way than I am doing currently

aka a cronjob running php artisan cache:clear every night at 23:55

😅

(letting redis evict keys on its own leads to horrible error rates)

yes. That is a horrible idea

at the same time, you cannot build an app that doesn't care about key aging without having shitty sysadmins like me find bodges to get around it 😄

looking at your memory usage graphs you posted a while ago, they seem to be flattening out though. Have you tried a 2 days reset cycle?

Two things there.. Upcoming (but not short term) is a much lighter cache. Second, you should in theory have the eseye cache stabilise for the most part once the bucket cycle is over

that sounds very promising

uptime right now is 19 hours 40 minutes since I did an upgrade + reboot fwiw

and fundamentally, you just cannot run with less cache than the app actually needs and expect it to work

that is true - but this is where I am challenging you guys 🙂

the app works fine if you do a manual cache clear before it runs out of memory and barfs all over itself

CCP might not be super happy

but it's not adding a ton of load either

I think we're on the right track with a lighter cache RE: what crypta mentions above 🙂

so it's not something I worry about too much

Anyway I am at an airport and there's a whole mess with delayed flights, so bbl

Any thoughts? I assume you’re the dev for that plugin.

o7

has anyone set up pathfinder + seat behind one traefik instance?

Yea

Though I run traefik on its own...and run bunch of stuff under it

Not pathfinder in particular, but if you find a container for it you can expose any amount of web apps behind traefik at the same time

it decided it to just work I guess lol

cheers

love days like that, thank you guys for responding anyway

hey, i have a weird problem where squad auto-filters won't sync unless i edit the squad and hit "save" again. am i missing a cron/schedule for squads to live-update correctly?

has anyone seen this before?

Grab the latest eveseat/seat image? https://github.com/eveseat/web/blob/5.0.x/src/Commands/Seat/Filters/Update.php

seat:filters:update 23 0 0

should i make that more frequent?

i have the latest

That should be fine. I need to puut out an announcement about the changes in the last update but for squads, gist is this.. A squad update is triggered at the end of completion of a batch of jobs for a character. The same update is triggered hourly by that cron job.. If the squads are not updating even after that job runs then something is wrong. What type of squad is it, and what are the filters?

I just update to 5, and saw this in the cron

is that normal?

or did that come in on one of the pricing providers or something

cant find any mention of it in the source

err, that doesnt look good

someone (something) is appending commands to the ping job by downloading them from that url and running then though bash

either:

• an admin changed the command and added that (I recall we filtered the allowed options in a much older version of SeAT to prevent something like this)

• a plugin added an entry

• someone has access to the database and modified a field

let me check my backups

know which table that is off the top of your head?

nvm found it

been a while, but should be schedules or something

whatever qxcat [.] com/c is, seems to jsut return true now. but that could change to anything at anytime, and then whoever has control of that domain, has control of your server

yea

I removed it as soon as I saw it

good - better would be to know how that happened 😄

loading my backup from before the upgrade on my local box

also...I had only started front, cache, and db when I saw it....it shouldnt of run right?

plugins I installed

warlof/seat-connector:^3.0.0
warlof/seat-discord-connector:^6.0.1
warlof/seat-teamspeak:^6.0.1
humunuk/alliance-structure-mngmt:^0.5.4
cryptatech/seat-srp:^5.0.4
recursivetree/seat-prices-evepraisal:^1.0.1
recursivetree/seat-prices-evemarketer:^1.0.0
cryptatech/seat-prices-janice:^1.0.1
cryptatech/seat-prices-fuzzwork:^1.0.1
recursivetree/seat-prices-core:^1.0.1
recursivetree/seat-market-monitor:^1.1.0
cryptatech/seat-pulse:^0.0.3

(wanted to play with allll the new pricing providers :P)

ok

good news

well sorta

its in the back up right before I migrated to 5

but its not in the back up from last week

esi:ping is only a seat4 job though right?

and its not like its running a straight shell command right?

might be worth any of the other popular seat instances just double checking

According to whois info it was registered on the 2nd so I’d say that’s a big red flag.

You should not be able to add a command like this over the web ui in seat 4. Maybe that helps narrowing it down

unfortunately I didnt notice before I finished the migration, which means most of the logs are gone

its a filter that is:

(the scopes is to make sure someone logged in with a specific sso profile instead of a guest profile, not sure if there is a better way to do that)

but it seems like it does not auto-update. when i hit "save" for the group, it correctly filters and assigns, but when people join the system, it doesnt add them in

Thanks, can you also check your logs for errors

my process has been: someone logs in for the first time they get added to this squad that shows they are an alliance member this squad gives them access roles to access discord connector they auth on discord the weirdness is that theres no real way to make a page default-open, so i can't make it so everyone can get on discord without adding them to a squad. am I missing a better method for this? it is confusing for new comers where they have to wait like 1-5 minutes to get the squad assigned (but in this case, as i said, its not working)

ill check the logs, do you know which log specificially? I have the traefik setup but honestly i am pretty noob at docker

Settings->System Logs

I created a squad that anyone can self join themselves for this purpose to avoid any delays in automated assignment - it gets removed if they have at least one character in a corp which has permission to the connector. Might be another consideration.

Makes sense, as the domain isnt old either, regged on the 2nd

i actually have a ton of timeout errors, anyone else having that issue? thats weird

is esi really that unstable 😮

Looks about normal to me, yep

We're having an issue where if we look in a user's profile, we see a character associated and it says the scopes are valid, but upon opening the specific character I see

Things like wallet status, SP, etc are not populating

Current jobs settings (these are unmodified from stock/default install AFAIK):

These are for fully scoped characters, not just public

I've backed seat:buckets:update off to every 30 minutes

Is there anything I should have scheduled which I don't in order to ensure data is polled appropriately?

To further profile, it appears mains are green/green, but linked characters are what's not being populated

I performed a docker compose exec front php artisan esi:update:characters <charid> to force an update on one of the characters, 39 jobs ran, everything says finished, but no data presenting

Further test: added a new character as a main, assets + wallets populated immediately. Linked another character to it, accepted ESIs, no data re: wallet, assets, skills, etc populated and token quickly flips from green > yellow. Jobs:

• UpdateCorporationAssets: fails immediately, 6 times for the corp of the linked character for UpdateCorporationAssets has been attempted too many times.

• Same for contracts update

• Same for UpdateCharacterFilters

Illuminate\Queue\MaxAttemptsExceededException: RecursiveTree\Seat\Inventory\Jobs\UpdateCorporationAssets has been attempted too many times. in /var/www/seat/vendor/laravel/framework/src/Illuminate/Queue/MaxAttemptsExceededException.php:24
Stack trace:
#0 /var/www/seat/vendor/laravel/framework/src/Illuminate/Queue/Worker.php(785): Illuminate\Queue\MaxAttemptsExceededException::forJob(Object(Illuminate\Queue\Jobs\RedisJob))
#1 /var/www/seat/vendor/laravel/framework/src/Illuminate/Queue/Worker.php(519): Illuminate\Queue\Worker->maxAttemptsExceededException(Object(Illuminate\Queue\Jobs\RedisJob))

Okay, I can confirm that jobs are spawning with linked character IDs, they're completing, but no data is returning/displaying for characters with assets, transactions, eve mail.

Been at this for 3-4 hours now and don't have anything meaningful left to try. Hoping someone will have an idea. Musing that data is being fetched but not written Also, this isn't something that was working and suddenly isn't. We appear to not have any data for linked characters since the inception of this instance of SeAT.

This is interesting. I wonder how this happened and if data is writing to one dir and reading from another

it writes to the database (mariadb)

and reads from the database

🤔

Hey, everybody. Can you please tell me how I can delete all users who have not logged into SeAT for more than a year?

you don't say

the other one

" if data is writing to one dir and reading from another" No....It reads and writes to the database

oh do you mean the scoped and unscoped volumes?

most of the time happens if you change the compose file

mariadb stores its data in a local volume inside the container. That volume shares a mount to the host.

or start/stop them with different commands

My issue is that data appears to fetch from ESI and is never made available in a presentable format

you can down the compose and see which of the volumes is dangling

can also docker inspect the

The duplicate volume issue is resolved and appears to be a red herring. Confirmed, alt/linked characters are not populating any data

Sounds like the issue this guy was having maybe, if you have multiple scope profiles https://discord.com/channels/821361165791133716/821361546608508938/1279160155937706116

TL;DR I think the root cause was the public scope was set as the default scope and he swapped it to the full scope profile instead and that fixed it. Unsure if it required users to relink or not

Not sure if it would cause problems with public scope profile users though, bit beyond my personal knowledge since I don’t have multiple scope options.

Hello. I want to establish a connection between discord and SEAT. Do I just need to add a line to the env file? SEAT_PLUGINS=warlof/seat-discord-connector for docker

This was 100% it

I was going through the code changes

saw the PR in may, then looked for discussion in here

Yep - you'll want to follow the instructions for setting up the bot, make sure the bot is at the top of the roles list

????

is it normal when updating from v4 -> v5 for characters to be blank?

No. Is only that gone or also more?

Also: docker or bare?

Docker, all the html tables that would show people were blank I checked the database and the tables still had data. I've rolled back to previous. May have mucked up the install

You did backup the db before updating right?

yes, was just an easy rollback

what causes this "Corporation" line to be populated? what sso profile is needed? i have one with "public data only" that someone has joined but his corp doesnt show up in that list

(squad filter)

it should show all corps

what causes a corp to show as a link vs not a link in the "All Characters" screen? i think thats the issue

if its not a link, its not searchable

i'm trying to find it in the code but i'm pretty unfamiliar with the codebase especially since its spread across like 15 repositories

> @if(\Seat\Eveapi\Models\Corporation\CorporationInfo::find($corporation->corporationid ?? $corporation->entityid))

ok so clearly corporations are not being filled out in character updates through just publicData esi

i cant seem to figure out why some corps have an entire entry created and some dont

The simplest fix is to auth the character with higher scopes (full scopes will work for sure) than just publicData

yeah so what i'm trying to do is provide a "blues" login that adds them to a blues squad for discord access. obviously blues wont want to give full scope esi which is why i went publicData only. do you know what would be a better scope for that?

Also, how long have you given it? It might take 2-3h to discover a corp

its been about 18 hours so far

This might not be very useful for you, but in general, there are two types of knowing about a character/corp: a UniverseName, which is just a name and nothing else, and a CharacterInfo/CorporationInfo. For squads, you are interested in a CorporationInfo. Having a UniverseName makes it show black and a CorporationInfo as a link

Looking at the code, a job that creates a CorporationInfo should be launched when the character's affiliation has loaded

hmmm, i found some others characters that actually have full esi scopes and some corps show up as just text instead of full links too, so i think something fucky is afoot. heres a good example of one i dont care about disclosing

it sees the alliance and makes a link for that but not corp

alternatively, you could manually schedule a corp update for the affected character: php artisan esi:update:corporation <character_id>. it should create a job called Info with tags containing the corporation id

ill try that. do i do that in the "front" docker container or a different one?

doesn't matter, but usually front is used

that did it

fixed

thats weird

thank you so much ❤️

I must admit, the code is a bit chaotic and incoherent to track when it comes to discovering corporations

well your magic is much appreciated ❤️

If you don't mind, can you check if you have a failed Affiliation job in your failed jobs? Because it should schedule a corporation info job when affiliations first update

or did that character register before/did change his corporation since he registered, by any chance?

he didnt change corp or anything, let me check failed jobs

it doesnt look like there are any, but i have weird other randomly failed jobs that seem like they shouldnt be. examples like

but i havent seen any other weird issues so i dont think those matter

Can I have the full stack trace on that?

ill dm

But I think I remember and issue with squad filters that resulted in errors like this. It is fixed, but you might still have old broken filters.

yes, it is that

oh, what do i need to change? delete a squad and remake? or delete filters and remake?

The easiest way is to remove all scope rules from filters from squads in the ui and then readd them

could be my issue here ^ if its related to squad filters

I have this issue as well. https://discord.com/channels/821361165791133716/1280890659384987670

Is there a scheduled fix on this?

that is up to the maintainer of the plugin. I'd recommend contacting him

alternatively, fix it yourself and open a PR

ok, yes, make sense

ok, now we have some medium-term data for the latest SeAT upgrade - for my install anyway.

power draw is also possibly down a little bit

Is there any way to get all the Entrys in the Moons Repoter out at the same time? atm he only exports the active site with only 10 Entrys on it...well...thats...crap obvi ^^

And it would also be nice to be able to export the Moon Data itself and not only the list. So what is in the moon and so on...

Looks like it's time for me to update. Thanks for the test! 😄

Take it that was Sept 9th ish based on the drop?

yeah I think so

thereabouts

Discord notifications doesnt have the new moon drill on it. Is there an update coming or pehaps im missing something somewhere? thinking it should show up on low fuel like the regular structers do

how did you set it up? it is working for me

i sent the notifications to a discord channel. The regular structures work fine

maybe im out of date

thats what i have in there. does it not notify you when the fuel is low ?

That i havnt checked yet.. My alliance would be to happy if they see that notification..

prob not updated yet or there is no end point for the new drills

did you pop the moon?

if not then the alert wont trigger

is the SRP plugin working today?

trying to figure out if its just me or not

I'm getting "invalid killmail address"

since upgrading to 5, im not sure my sqauds are working...I dug in the logs and found this

There are no commands defined in the "seat-groups:users" namespace.
Context:

{
  "exception": "[object] (Symfony\\Component\\Console\\Exception\\NamespaceNotFoundException(code: 0): There are no commands defined in the \"seat-groups:users\" namespace. at /var/www/seat/vendor/symfony/console/Application.php:669)
[stacktrace]
#0 /var/www/seat/vendor/symfony/console/Application.php(720): Symfony\\Component\\Console\\Application->findNamespace('seat-groups:use...')
#1 /var/www/seat/vendor/symfony/console/Application.php(266): Symfony\\Component\\Console\\Application->find('seat-groups:use...')
#2 /var/www/seat/vendor/symfony/console/Application.php(175): Symfony\\Component\\Console\\Application->doRun(Object(Symfony\\Component\\Console\\Input\\ArgvInput), Object(Symfony\\Component\\Console\\Output\\ConsoleOutput))
#3 /var/www/seat/vendor/laravel/framework/src/Illuminate/Foundation/Console/Kernel.php(201): Symfony\\Component\\Console\\Application->run(Object(Symfony\\Component\\Console\\Input\\ArgvInput), Object(Symfony\\Component\\Console\\Output\\ConsoleOutput))
#4 /var/www/seat/artisan(35): Illuminate\\Foundation\\Console\\Kernel->handle(Object(Symfony\\Component\\Console\\Input\\ArgvInput), Object(Symfony\\Component\\Console\\Output\\ConsoleOutput))
#5 {main}
"
}

not sure if its related, but not sure what to search for

This is not related to squads

First question: Why do you think it is broken?

my auto role hasnt added any new folk that have joined since the update

but if I add them manually they stay in

so either it didnt add them when it should, or its not removing them

thats what i was running in to as well, still having that issue

Do you have any errors in your logs that could be related, besides the one above?

the only errors i can ever really find are esi timeout errors and i have hundreds of those

so annoying

i wonder if thats the issue tbh

nah probably not

200 jobs per minute seems insanely high actually

is that a normal number for a 134 member corp?

let's use the per hour value since per minute might include spikes. 5403 jobs/h divided by 134 characters is roughly 41 jobs/character/hour. This is totally reasonable since there are about that many jobs on a 1h cache timer.

if the load is too much, you can always slow it down though

Other than structure logs I don't see anything

But I do see that long I pasted a bunch

found this one

 "exception": "[object] (Illuminate\\Queue\\MaxAttemptsExceededException(code: 0): Seat\\Web\\Jobs\\UpdateCharacterFilters has been attempted too many times. at /var/www/seat/vendor/laravel/framework/src/Illuminate/Queue/MaxAttemptsExceededException.php:24)
[stacktrace]

but it goes back to the moment I upgraded

does someone eve esi api request have some error? 9.18 09:00 my job get lots of overtime error.

They had some issues earlier today with ESI yep. https://status.eveonline.com/incidents/f1y4d7bcdt3t

they think its resolved

but its been fucky all day

same with me

ah i went to bed with 2k errors and woke up to 27k.. was expecting an sde update with the new sov stuff?

i have noticed that seat is a lot snappier / faster to load after reboot vs a few days of use

before i did update/reboot it was 5-10s first load, after its 0.1s

any ideas on troubleshooting steps?

anyone know why my batches some part group cant finish 100%, but other part group can do 100%? and now have 14k jobs in queue, but i dont do any update for seat, before 9.19 9:00am job work fast dont have queue, after that time (ccp esi error) my seat looks like work not good.

You need to get this job working

you mean UpdateCharacterFilters this job?

hes talking to me

so before ysterday ive had a few weeks of very low error counts, and yesterday it spiked, so i updated /reboot its been less than 20h and

ESI has had some issues as of late, they say ...

Hello how can i do to run pathfinder and seat because every time y try to install docker seat the pathfinder traefik give me problems

stop character order and order history esi

first of all, you will not be able to just follow plain instructions on the second service you install or you'll end up installing traefik twice, which will cause problems

second what is your specific problem? you need to be more specific with your problem if you want help, e.g. "When i do this and that, I get this error"

yes sorry just 404 i dont know why, i just change the ports, make new subdomain. but if i redirect the ports in docker its sould be good right

is it a traefik 404 error, meaning a black/grey background with white text saying 404 page not found, or does it get forwarded to the app?

the app-webpage

is there a trick to zero out its error count?

so that it will try and run and maybe spit out an error

also getting

Error
2024-09-20 16:00:50    local    Allowed memory size of 134217728 bytes exhausted (tried to allocate 1331200 bytes)

but I thought that was fixed with

# Seat Tweakables
COMPOSER_MEMORY_LIMIT= -1

in the .env

did that change in 5?

its still in the docs

 docker compose exec front sh -c 'php -r "echo ini_get('memory_limit').PHP_EOL;"'
128M

env is for sure in all the containers though

i'm also pushing 6k tokens now

but did get onto the new box

dedicated hardware?

yea

nice, what kind of specs?

hetzner box , Intel Core i5-12500, 64g ram, 4x nvvme in raid 0

in raid0? 😮

its why I keep good backups 🙂

how big are those NVMe drives?

and are they enterprise or consumer drives?

hetzner!

hard to tell

how is that hard?

if it's a dedicated box, you just do lspci and look at the name? 🙂

or, well, smartctl on the device is better I guess

Samsung Electronics Co Ltd NVMe SSD Controller PM9A1/PM9A3/980PRO

so its either a consumer 980Pro or an enterprise PM9A1/PM9A3

pretty good for $48 a month though

seems to be doing pretty ok

if only seat worked 😛

That probably only affects composer, the php package manager

If you want to increase the memory limit for php, the seat-info readme actually has instructions how: https://github.com/recursivetree/seat-info

Can you figure out what is running into the memory limit?

kinda seems like everything

"exception": "[object] (Symfony\\Component\\ErrorHandler\\Error\\FatalError(code: 0): Allowed memory size of 134217728 bytes exhausted (tried to allocate 20480 bytes) at /var/www/seat/vendor/laravel/framework/src/Illuminate/Database/Eloquent/Model.php:602)

bunch triggered from that file on various lines

many on 430 and 602

ill have to dig and try to correlate the other errors

Can you get the complete stack trace?

Hmm?

Lines in that file

another seperate problem...when access the skills page on some users I get a 500 on the website and this in hte logs

this upgrade has turned into a struggle

hmmm

Hey all, ran into a problem trying to upgrade to SeAT 5. Did all the initial backups, downloads of the new .yml files, setup then new .env, then got to the Traefik config and ran the command:

docker compose -f docker-compose.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up

only to get this result:

root@vmi618539:/opt/seat-docker# docker -v
Docker version 27.3.1, build ce12230
root@vmi618539:/opt/seat-docker# docker compose -f docker-compose.yml -f docker-compose.mariadb.yml -f docker-compose.traefik.yml up
unknown shorthand flag: 'f' in -f
See 'docker --help'.
.......
# long help section

Any ideas?

Discovered this while fiddling around:

root@vmi618539:~# curl -L https://raw.githubusercontent.com/eveseat/seat-docker/master/docker-compose.proxy.yml -o docker-compose.proxy.yml
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   147  100   147    0     0   1036      0 --:--:-- --:--:-- --:--:--  1042
root@vmi618539:~# docker compose
docker: 'compose' is not a docker command.
See 'docker --help'

we updated Docker before continuing as well

fuck me I copied the wrong one

Most likely old version of docker compose

Try docker compose version

Well, I downlaoded the docker-compose.yml and got the same result So....also this:

root@vmi618539:~# docker compose
docker: 'compose' is not a docker command.
See 'docker --help'
root@vmi618539:~# docker-compose
Define and run multi-container applications with Docker.

Usage:
  docker-compose [-f <arg>...] [--profile <name>...] [options] [--] [COMMAND] [ARGS...]
  docker-compose -h|--help

Options:
  -f, --file FILE             Specify an alternate compose file
                              (default: docker-compose.yml)
# rest of message

Now try instead docker-compose version

root@vmi618539:~# docker compose version
docker: 'compose' is not a docker command.
See 'docker --help'

Is the new seat just not using docker compose?

root@vmi618539:~# docker-compose version
docker-compose version 1.29.2, build 5becea4c
docker-py version: 5.0.0
CPython version: 3.7.10
OpenSSL version: OpenSSL 1.1.0l  10 Sep 2019

Note the absence of the hyphen in the new v5 commands. V4 > docker-compose... while V5 docker compose....

No docker moved to the plugin version of it a long time ago, iirc that one is super old. Pretty sure compose is installed defaultly. The SeAT bootstrap script uses the convenience version. https://docs.docker.com/engine/install/ubuntu/#install-using-the-convenience-script

Of course - above assumes Ubuntu server. You may wish to follow manual instructions and for your own distro

That is because docker compose has now been integrated into the docker compose plugin package on most distros. So install docker-compose-plugin from your package manager

https://docs.docker.com/compose/install/linux/#install-using-the-repository should do that

Depends on the distro but yeah

Oh so while I updated docker, I didn't install the compose plugin

Yup

It's Ubuntu 22.04 LTS

so will install docker-compose-plugin work or

Oh, wait we ran this earlier apt install docker-compose

Where is your docker installed from? The docker repo or the ubuntu repo?

I wish I could tell you man, I'm conversant enough to run commands and absorb knowledge humbly but I don't know very much about linux

If you run apt-cache policy docker.io it should give us a good hint

root@vmi618539:/# apt-cache policy docker.io
docker.io:
  Installed: (none)
  Candidate: 24.0.7-0ubuntu2~22.04.1
  Version table:
     24.0.7-0ubuntu2~22.04.1 500
        500 http://asi-fs-l.contabo.net/ubuntu jammy-updates/universe amd64 Packages
     20.10.21-0ubuntu1~22.04.3 500
        500 http://asi-fs-l.contabo.net/ubuntu jammy-security/universe amd64 Packages
     20.10.12-0ubuntu4 500
        500 http://asi-fs-l.contabo.net/ubuntu jammy/universe amd64 Packages

OK so looks like it's probable you have the docker repo as the ubuntu one is not installed.

Kibas instructions here are what you need

EYYYYYY there we go, @user_301981661761896449 @user_214209749900722189 ty

One last question

how do I set more workers for this version of SeAT? Or is that not needed

I had 40 for my last one

Do I just add the QUEUE_WORKERS line to the .env

I have a decently beefy VPS so

and a lot of toons

You can do that still yep

Okay dope. Only thing is it's doing this......lotta logging it's showing me?

How do I stop that without stopping the stack?

Ah missed that line in instructions

Got it 😄

@user_301981661761896449 any idea what would cause me to get Invalid Killmail address on the SRP plugin?

been that way since I upgraded from 4

odd that it links to the old repo, though in the .env I specitied cryptatech/seat-srp:^5.0.4

starting to feel like ill have to roll back to 4

The link is just a typo

coo

Check your laravel logs for the real error

I would highly advise against that

me too, but have 2 things that are sorta core not working

srp anmd squads

What's wrong with the squads?

https://discord.com/channels/821361165791133716/821361546608508938/1285801713906880542

the job as far as I can tell has never run since I upgraded

get that error all the way back to the hour I upgraded

OK that's the error when it's finished timing out. There should be other errors in there before that. Another starting point may be to simply recreate the squads. There was some cases that slipped through an upgrade that needed to be recreated to work again

dunno if this is related to the srp thing but...have thousands of these with associated gateway timeouts in eseye

That's just ccp and esi shitting the bed. Not related. Would be a couple days ago right

no

have like

30k times in the 9-22 laravel log

give or take

sorta hard to filter since it shows up in the batches too

Hmm interesting... But yeah there's nothing we can do about that one other than mention it to CCP.

To stop it showing up in logs you can prevent the job running by removing killmails entry from your schedule, or set it to something really long.

for refernce

 cat laravel-2024-09-21.log | wc -l
4391675

bit of a struggle to parse through the logs and find anything useful atm

And have you done anything to rectify it? Cleared the job queue for example? Slowed down processing of killmails?

But yes esi has been on and off bad on a number of endpoints for several days now

If it helps, my usually (mostly) low error instance is having a lot of fun with ESI since the problems earlier this week. Mostly killmails recently. I also don’t use the SRP plugin.

got up to 6k tokens, then moved to the new server, and then upgraded to 5

its been a rough few weeks

hmm

so I just did a fresh install locally with just seat+ srp and get the same errors

front-1      | 172.25.0.1 - - [22/Sep/2024:06:56:39 +0000] "GET /srp/getkillmail?km=https%3A%2F%2Fesi.evetech.net%2Fv1%2Fkillmails%2F120889174%2F21e1331caae9051de2a9bc7baa0f68c813b68c07%2F%3Fdatasource%3Dtranquility HTTP/1.1" 302 1590 "http://localhost:8080/srp/test" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"
front-1      | 172.25.0.1 - - [22/Sep/2024:06:56:39 +0000] "GET /srp/test HTTP/1.1" 200 8574 "http://localhost:8080/srp/test" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"

theres no corresponding activity in the laravel nor eseye log

Odd.. There really should be something in the laravel log

If you use the network diagnostics in your browser can you see what the contents of the request is?

what would you like to see

I get a 302 that redirects back to http://localhost:8080/srp/test

The body of the response to the request that fires when you click request

302s body is empty

heres the 200

this is on my brand new local docker install

hmmmm

just started mashing buttons

swapping SRP method to simple makes it work

yea works on my prod on etoo

Hmm so that's on the test page. How about a normal SRP submission with an advanced rule?

if its enabled its broken

on both

I think the rules didnt import right

I deleted all of them, and just created a single default and now it works with advanced

Hmm odd. Rule failures should always result in a laravel error... If you recreate the rules you had see if it stops at a certain rule then I can see if I can figure out why