anand388[m] has quit [Quit: Idle timeout reached: 10800s]
<VITAS[m]>
how are things?
<VITAS[m]>
and stuff?
RockyTV[m] has quit [Quit: Idle timeout reached: 10800s]
VITAS[m] has quit [Quit: Idle timeout reached: 10800s]
RockyTV[m] has joined #spacedock
<RockyTV[m]>
was that for me? <span class="d-mention d-user">VITAS</span>
VITAS[m] has joined #spacedock
<VITAS[m]>
yes
NibySmurf[m] has joined #spacedock
<NibySmurf[m]>
Could anyone help me to get ScienceAlert ReAlerted working?
DasSkelett[m] has joined #spacedock
<DasSkelett[m]>
The people on the forum thread probably can: https://forum.kerbalspaceprogram.com/index.php?/topic/175772-19x-sciencealert-realerted-experiment-availability-now-with-dmagic-support/
<RockyTV[m]>
<span class="d-mention d-user">VITAS</span> not much, still getting used to work. I was supposed to test something last week but the feature wasn't ready so now I'm just watching my friend that works at the same company stream what he's doing so I can pickup some stuff and understand what's he doing
<VITAS[m]>
watching streams at work :D
<VITAS[m]>
why is this channel suddenly used by endusers?
Darklight[m] has joined #spacedock
<Darklight[m]>
shrug but we need to figure out how to delegate and find where the modders live
<Darklight[m]>
I lost contact with other modders
<VITAS[m]>
i didnt but we cant just send peopel to some internal channels
<VITAS[m]>
so i guess forums?
NibySmurf[m] has quit [Quit: Idle timeout reached: 10800s]
DasSkelett[m] has quit [Quit: Idle timeout reached: 10800s]
RockyTV[m] has quit [Quit: Idle timeout reached: 10800s]
<VITAS[m]>
can someone verify that?
<VITAS[m]>
Just an FYI, I tried uploading to Spacedock,and it's just hanging after the bar gets to 100%
DasSkelett[m] has joined #spacedock
<DasSkelett[m]>
Can someone verify what?
<VITAS[m]>
Just an FYI, I tried uploading to Spacedock,and it's just hanging after the bar gets to 100%
Geo[m] has joined #spacedock
<Geo[m]>
I can't make a spacedock account...
<DasSkelett[m]>
Oh now I understand
<DasSkelett[m]>
I'll check
<VITAS[m]>
Geo: where does it hang?
<VITAS[m]>
DasSkelett: Linuxgurugamer just send that on the forums to me
<DasSkelett[m]>
Updating works fine.
<VITAS[m]>
sec im telling him to come here
<Geo[m]>
like I can't verify
<VITAS[m]>
so no email?
<VITAS[m]>
DasSkelett: i send Linuxgurugamer a message
<VITAS[m]>
Geo? no email?
<Geo[m]>
no
<Geo[m]>
and I tried sending the verification email as well
<Geo[m]>
the bot command
RockyTV[m] has joined #spacedock
<RockyTV[m]>
check spam folder
<VITAS[m]>
whats your username?
<Geo[m]>
geoplex_official
<VITAS[m]>
cant se anyone by this username in the db
<DasSkelett[m]>
Is the db query case sensitive? The admin interface says `Geoplex_Official`
<Linuxgurugamer[m]>
<span class="d-mention d-user">VITAS</span> good morning
<VITAS[m]>
morning
<VITAS[m]>
you have a problem?
<VITAS[m]>
DasSkelett: couldnt replicate it
<Linuxgurugamer[m]>
hi
<Linuxgurugamer[m]>
sorry, working. Not sure, may have been my system
<Linuxgurugamer[m]>
I did an upload, and it just hung when the upload was complete
<Linuxgurugamer[m]>
but then I noticed that it had actually done it
<Linuxgurugamer[m]>
But now I'm wondering about the CKAN notification, the mod has not been updated in CKAN yet (achievements)
<DasSkelett[m]>
Taking a look
<Linuxgurugamer[m]>
ok. mod # 2422
<Linuxgurugamer[m]>
and I just confirmed that CKAN gets it'
<DasSkelett[m]>
Thanks
<Linuxgurugamer[m]>
*edit:* ~~and I just confirmed that CKAN gets it'~~ -> and I just confirmed that CKAN gets it's updates from Spacedock for this one
<Linuxgurugamer[m]>
I'm checking the ckan-meta right now
<Linuxgurugamer[m]>
ok, confirmed that the ckan meta was not updated to 1.10.0.3
<Linuxgurugamer[m]>
thank you
<VITAS[m]>
:)
<DasSkelett[m]>
Hmm yeah, nothing there yet. I'll wait a bit and see if the webhook will still trigger, sometimes the it takes some time.
<Linuxgurugamer[m]>
ok.
<VITAS[m]>
i restarted prod after that message
<VITAS[m]>
just in case
<Linuxgurugamer[m]>
it updated
<Linuxgurugamer[m]>
I see the new file in ckan-meta
<Linuxgurugamer[m]>
so I would expect that theclients should see it soon
<DasSkelett[m]>
Yep, can see it in CKAN 👍
<Linuxgurugamer[m]>
your local ckan?
<DasSkelett[m]>
Yes, the client
<Linuxgurugamer[m]>
ah. my was slow for some reason, just showed up. Ok, thank you
<DasSkelett[m]>
Yeah I found out that GitHub needs a few minutes to update the zip download or repositories.
<DasSkelett[m]>
I think they are handled differently than the commits themselves, because commits are updated basically in real-time, but zip downloads not. Maybe it needs some time to propagate through their datacenters.
<Linuxgurugamer[m]>
probably. Anyway, I have to get back to my day job
<DasSkelett[m]>
👋
Darklight[m] has quit [Quit: Idle timeout reached: 10800s]
HebaruSan[m] has joined #spacedock
<HebaruSan[m]>
Hmm, the SpaceDock API JSON format puts unnecessary spaces between fields, minor but might be worth it to remove those
Darklight[m] has joined #spacedock
<Darklight[m]>
VITAS I figured out how to get the thing to stop going read-only
<Darklight[m]>
At least in my local testing... the solution to me is unacceptable... ``tune2fs -O ^has_journal /dev/loop0``
<Darklight[m]>
The error is within ext4 itself, when the journal craps out ext4 goes "nope" and it doesn't matter what mount options you use, errors=continue does not stop it from going RO, it's related to the journal
<Darklight[m]>
I accidently figured that out after not experiencing this on ext2 which lacks a journal
<VITAS[m]>
so explanation but not solution
<VITAS[m]>
result: go to bed we try again tomorrow
<Darklight[m]>
If it was me I'd accept not using quotas and using subvol
<Darklight[m]>
Network vdisk already sketches me out
<Darklight[m]>
I don't know why nbd worked exactly correctly but I can't shake the feeling it's dodgy
<Darklight[m]>
The nbd approach does have a few advantages, it brings back quotas, it maintains *one* smb connection and one only, for some reason the writes actually go through when smb comes back up but I don't understand how...
<VITAS[m]>
i will try till we do the next prod update and use the best solution we have by then
<Darklight[m]>
The changes to the filesystem seem to sit in ram and still can be written to
<Darklight[m]>
When smb comes back up it seems like linux is smart enough to actually write to disk
<Darklight[m]>
I didn't get disk corruption with no journal, but I think you can bet your bottom dollar it will eventually happen... I'd still be poking that file server and dumping encrypted storage on it 😛
<Darklight[m]>
Unless you have some other type of backup
<VITAS[m]>
file backup?
<VITAS[m]>
yes i do those
RockyTV[m] has quit [Quit: Idle timeout reached: 10800s]
Geo[m] has quit [Quit: Idle timeout reached: 10800s]
<HebaruSan[m]>
How do you want to handle migrating code from beta to prod/master this time? Any chance we could merge it on the GitHub remote first then pull to prod?
<VITAS[m]>
we have db schema changes ?
<DasSkelett[m]>
Yes
<VITAS[m]>
I wills etup a second container, we deploy the code onto it and keep it mapped to beta db.
<VITAS[m]>
when its working like it should we can schedule a downtime for prod and update the db schema and switch over to the new instance in the rev proxy
<VITAS[m]>
after that we can bring the site up again and update the other prod container
<HebaruSan[m]>
But we're not going to randomly pull GitHub's beta branch to prod and leave master unupdated again, right?
<VITAS[m]>
nope
<HebaruSan[m]>
Cool, cool
<VITAS[m]>
i want to be sure that we have fixed procedures
<DasSkelett[m]>
If we want to do both these things, we have to disable the auto-update webhook beforehand.
<HebaruSan[m]>
I don't think the auto update is active in prod
<DasSkelett[m]>
Okay
<DasSkelett[m]>
We should re-check though, to be sure.
<VITAS[m]>
its also likely that we will change mod storage on the os level
<VITAS[m]>
so it will be neccesary to copy all mods to a different location
<VITAS[m]>
that will take some time
<HebaruSan[m]>
From a contributor point of view it would be great if pushing to master would update prod, but I understand that VITAS wants more control than that
<VITAS[m]>
and ive to figure out how to make them available to both containers at the same time
<VITAS[m]>
yes i want well planed and announced downtimes
<VITAS[m]>
we could add a "staging" branch
<DasSkelett[m]>
Another one?
<VITAS[m]>
staging would mean that we deploy the code onto the hotspare container in the future
<VITAS[m]>
if it helps HebaruSan
<HebaruSan[m]>
Frankly I think I would prefer fewer branches, not more
<DasSkelett[m]>
Yep, same
<HebaruSan[m]>
alpha and beta are already somewhat redundant
<VITAS[m]>
i personaly would liek the code on prod and prodspare to be the same unless an iminent update is performed
<VITAS[m]>
k
<VITAS[m]>
how about adding some info to git when we updated prod last?
<VITAS[m]>
like an issue or something?
<VITAS[m]>
we could see what code was added after it and issues are resloved since then
<DasSkelett[m]>
For better readability pass it through `jq` or use a proper browser that does it for you. It doesn't have to be beautiful, it's an _Application_ Programmin Interface after all. For humans there is a nice frontend to use.
<DasSkelett[m]>
For better readability pass it through `jq` or use a proper browser that does it for you. It doesn't have to be beautiful, it's an _Application_ Programmin Interface after all. For humans there is a nice frontend to use.
<HebaruSan[m]>
Or pipe it to `python -m json.tool`
<DasSkelett[m]>
`jq` is shorter and has a coloured output 😛
<HebaruSan[m]>
Ooh, nice
<HebaruSan[m]>
I usually only use jq when I need to do processing
<DasSkelett[m]>
Yeah I started to use it also if I just need to make some JSON better readable. It works excellently, has color output, and is short to type: `curl http://localhost:5080/api/mod/1 | jq`
<DasSkelett[m]>
Hmm, interesting. Last time I checked https://spacedock.info/random it was very slow (`/browse/top`-level slow), I thought it was because it loaded all mods into an array first, and the chose one.
<DasSkelett[m]>
However testing it now, it's really fast. Wonder what happened there.