Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2018-01-03

| Channels | #salt index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 saltslackbridge joined #salt
00:15 mikecmpb_ joined #salt
00:27 mikecmpbll joined #salt
00:32 mikecmpb_ joined #salt
00:39 evle3 joined #salt
00:41 pipps joined #salt
00:52 bildz joined #salt
00:53 evle1 joined #salt
00:55 miruoy joined #salt
00:59 evle2 joined #salt
01:15 evle joined #salt
01:22 edrocks joined #salt
01:31 aldevar joined #salt
01:37 major joined #salt
02:25 vexati0n salt-cloud is like 75% weird issues
02:34 nomeed joined #salt
02:38 Deliant joined #salt
02:59 ilbot3 joined #salt
02:59 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2016.11.8, 2017.7.2 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic, and https://saltstackcommunity.herokuapp.com (for slack) <+> We are volunteers and may not have immediate answers
03:37 MTecknology heheh... maybe only 25%, but those issues tend to stand out
03:55 onlyanegg joined #salt
04:21 masber joined #salt
04:38 nethershaw joined #salt
04:41 justanotheruser joined #salt
04:44 chowmeined joined #salt
04:52 indistylo joined #salt
05:23 zer0def joined #salt
05:23 onlyanegg joined #salt
05:29 shiranaihito joined #salt
05:47 onlyanegg joined #salt
05:54 girishb joined #salt
05:57 indistylo joined #salt
05:58 onlyanegg joined #salt
05:59 aruns joined #salt
06:03 aruns__ joined #salt
06:08 tflo joined #salt
06:09 tflo left #salt
06:18 LocaMocha joined #salt
06:31 tflo_ joined #salt
06:32 aruns joined #salt
06:53 aruns__ joined #salt
06:59 onlyanegg joined #salt
07:48 pualj joined #salt
07:48 Tucky joined #salt
07:55 nethershaw joined #salt
08:13 girishb joined #salt
08:13 Hybrid joined #salt
08:26 ech042 joined #salt
08:54 aruns joined #salt
09:00 onlyanegg joined #salt
09:04 aldevar joined #salt
09:06 mavhq joined #salt
09:07 girishb joined #salt
09:07 mikecmpbll joined #salt
09:19 aldevar joined #salt
09:20 marcinkuzminski_ joined #salt
09:55 Hybrid joined #salt
10:18 willprice94 joined #salt
10:18 willprice joined #salt
10:22 evle joined #salt
10:26 pualj joined #salt
10:26 mschiff joined #salt
10:43 colegatron joined #salt
10:43 colegatron morning all
10:47 colegatron I am trying to figure out the best way to get, in a pillar, the ethernet IP of the host where the state using that pillar is running. salt["grains.item"](ip4_interfaces:eth0"])[1] is not fine, because some servers has 'eth0', some 'esn3', and probably some others
10:49 colegatron I've followed the approach of iterate over the ip4_interfaces not 'lo' and get the first ip in the list, but what I wrote is not working and not sure how to display text in the logs to debug it
10:49 colegatron here is the code: https://pastebin.com/qbdeyDnW
10:50 colegatron there is any other approach? or at least a way to show things on the logs?
10:50 colegatron tnx
10:54 haam3r_ colegatron: I've just used something like this: {% set ip = salt['grains.get']('ipv4')[0] %}
10:55 colegatron I tried, but you can end up with a 127.0.0.1
10:55 haam3r_ yeah, just saw that for the first time too
10:55 colegatron or the docker interface ip (if it is already installed when you run the state)
11:00 haam3r_ do you need the private address or could it also be public?
11:01 onlyanegg joined #salt
11:01 colegatron should be the private one, in fact there will be just one assigned to the iface. i'm using aws
11:09 babilen colegatron: Why not use network.ip_addrs ? You can give it a CIDR or type={public,private} and you don't have to hardcode interfaces
11:10 babilen Or do you really care about interfaces?
11:11 colegatron I need the ethernet one. docker swarm needs advertise the eth ip.
11:11 babilen Do you know the CIDR?
11:11 babilen Or rather: What's the invariant?
11:11 colegatron well, yes, it would be a 172.31.0.0/16
11:12 babilen So use that CIDR (might want to pass it in via pillars)
11:13 babilen My approach is to - generally - create a mine_function alias for every CIDR with some semantics associated
11:13 colegatron ahh, I see the doc. nice. that's the best solution.
11:13 babilen That allows me to also query that information for other minions
11:15 colegatron good. thank you very much, didn't know about the cidr argument in the network module
11:17 babilen http://paste.debian.net/1003384/ is what I use as mine function alias (and so on for other CIDRS with a meaning)
11:18 babilen You could probably grab the list of CIDRs via boto and generate that automagically
11:19 colegatron umm never used mine functions (mine in general either)
11:20 babilen {% set nfs_server_ip = salt['mine.get'](nfs_server.tgt, nfs_server.addr_mine_function).values()[0][0] %}
11:20 babilen That's what I use to get the address of a NFS server (I pass in the target expression and specific mine function via pillars
11:21 babilen That way I can easily say "The NFS server is the one with foo-bar* ID and the network is foo_private_bar"
11:22 babilen nfs_server is the pillar with the NFS settings (for example)
11:22 colegatron salt pro :)
11:23 babilen It comes in handy if you need this information from other minions also .. and I rather have one way for all minions to get addresses
11:23 sils joined #salt
11:23 babilen It really isn't that hard .. I think the tricky bit was figuring out how to piece it together ..
11:24 colegatron sure, everything is easy one you grap it all :-P
11:24 colegatron grasp
11:24 babilen I had problems with hardcoded interfaces once and just said: "No, we're not doing that again. I don't care about interfaces for most setups, so lets use networks (i.e. cidr) insteat"
11:26 colegatron joined #salt
11:27 colegatron Do you know if the module function can be called within the pillar itself? (and if the resulting IP will be the master or the minion one)
11:28 colegatron not sure who parses(and runs) the pillar file before being used by an state
11:28 babilen Which module?
11:29 babilen I use mine.get in the pillar as documented on https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mine.html#salt.modules.mine.get -- see the runner bit
11:29 babilen (regardless if this is a local or a remote address)
11:29 colegatron network.ip_addrs(cidr='172.16.0.0/16)
11:29 babilen I guess you can do the same with network.ip_addrs
11:30 colegatron you use it though a function file
11:30 colegatron and I meant avoid it and run it directly in the pillar
11:30 babilen What I do is to target the mine function aliases (see above) to all/suitable minions, refresh the mine and use mine.get to get addresses everywhere else (SLS files, pillars, ...)
11:31 colegatron oh sorry, I see; salt["mine.get"]
11:32 babilen Unfortunately there isn't a network runner, so that runner approach doesn't work for you
11:33 babilen Which means that if you use the network.ip_addrs execution module in a pillar it will be executed in the master's context
11:33 babilen https://docs.saltstack.com/en/latest/ref/runners/all/index.html#all-salt-runners
11:34 mschiff joined #salt
11:35 babilen Looks as if using the mine for this is a sensible approach
11:37 colegatron no, not really. just not used to. to be honest I'm using salt for two years now and I am just reading what a runner is. :-(
11:38 babilen Hmm .. I wonder if there's a high level "these are the major components and how they relate to each other" document available these days
11:41 colegatron that's the document I always expected to find somewhere when I was starting with salt :-P Once everything was working never had to dig deeper (and expected to not have to due the confusion with terms and features). not sure if I'm too lazy or too dumb :)
11:42 mikecmpbll joined #salt
11:46 ventris joined #salt
11:51 ventris left #salt
11:54 babilen colegatron: Had a short look and didn't find anything like that
11:57 colegatron thank you babilen
11:59 babilen I mean I'm constantly finding new things ... Just the other day I stumbled over the @jinja_filter(...) decorator that isn't documented anywhere, but that allows you to define your own jinja filters easily in utils/... modules
11:59 babilen That way you can incorporate Python code quite easily in Jinja with |call_some_function()
12:00 babilen https://docs.saltstack.com/en/latest/topics/jinja/index.html#filters are the built in ones, but you can easily define more
12:00 babilen That makes jinja a lot more usable
12:01 sils joined #salt
12:04 colegatron sure. the thing is that I don't really need to handle too much infrastructure (<50 servers) and it is pretty static and I hadn't the need to dive so deep into salt to have everything as I needed. It is always nice/funny have something to dig on but now I am starting to move to docker and maybe salt (or any other cfg mmgt) is not going to be very useful, aside of adding another abstraction layer to the infrastructure.
12:04 babilen kubernetes?
12:04 colegatron nope. swarm
12:05 babilen We have some workloads on k8s and I'm quite happy with it, but I definitely need to learn more
12:06 colegatron looked into k8s but like to stick to the kiss principle, so I choosed swarm. it is mature and will be getting better with the time, I think
12:07 babilen My impression is that k8s essentially won
12:07 babilen But I won't evangelise :)
12:08 colegatron yep. it was the first, and it is from G, but there are other players. same happens with Cheff,Puppet and salt. I and choosed salt ;-)
12:19 gmoro joined #salt
12:33 cyteen joined #salt
12:34 aldevar joined #salt
12:38 zer0def anyone looked into AWS ACM management from Salt?
12:40 zer0def ACM, as in their own Let's Encrypt
12:52 aruns__ joined #salt
13:01 onlyanegg joined #salt
13:14 Nahual joined #salt
13:22 sjorge joined #salt
13:49 aldevar joined #salt
13:53 pipps joined #salt
13:54 pipps joined #salt
14:01 CrummyGummy joined #salt
14:02 izibi joined #salt
14:05 edrocks joined #salt
14:14 dlloyd hi all, has anyone else noticed the postgres.create_user slowing geometrically as the count of users in a db grows? in debug mode i see it iterating every user in the db after each create statement, so when i create 200+ users it... takes some time
14:20 babilen I guess the best way to address that issue would be to introduce a "users: ..." argument to that state that allows you to specify multiple users per state
14:20 babilen That way it only needs to iterate once
14:21 babilen How long does this take? I would have hoped that it runs rather fast (I mean its one query and a filter)
14:22 babilen Or it really should be one query, shouldn't it? (select from role table)
14:24 dlloyd it looks like its the list users part of it, has to get info on each user
14:24 babilen God, that's horribly implemented :(
14:24 babilen role_get should be a single query
14:24 dlloyd so it runs that on each user, after each create, so growing
14:25 dlloyd and not fast, on a local 9.4 instance, running against a ram disk, its going on 30 minutes to create ~230 users
14:25 babilen https://github.com/saltstack/salt/blob/develop/salt/modules/postgres.py#L942-L954
14:25 babilen 30 MINUTES?
14:25 dlloyd mhmm
14:26 dlloyd hold on, can generate output for a simple 5 user run
14:26 babilen https://github.com/saltstack/salt/blob/develop/salt/modules/postgres.py#L897-L926
14:26 babilen Holy shit ... That needs urgent refactoring
14:27 babilen I mean, that should be two queries per "add user" not all that madness
14:27 babilen And you should even be able to short circuit it when the user doesn't exist (you don't need the inherited roles and whatnot for a role that doesn't exist in the first place)
14:29 babilen Please file a bug, recommend a "user_exists" function in the module that runs for a specific user (SELECT 1 FROM pg_roles WHERE rolname=$USERNAME) and only if it exists starts getting info for that specific user and not *ALL OF THEM*
14:29 babilen Sorry, can't be of more help short of implementing that myself
14:29 dlloyd will do
14:31 babilen https://github.com/ansible/ansible/blob/devel/lib/ansible/modules/database/postgresql/postgresql_user.py#L245-L251 -- Ansible does this
14:32 babilen That Ansible module looks a lot nicer (proper use of itertools, frozensets, ...) than the SaltStack equivalent
14:39 girishb joined #salt
14:47 cgiroua joined #salt
14:57 concerti joined #salt
15:02 onlyanegg joined #salt
15:12 pualj joined #salt
15:15 XenophonF zer0def: I've looked at using ACM from Salt, but like Let's Encrypt is isn't something that can easily be used in a state
15:22 XenophonF is it possible to use ZeroMQ and RAET at the same time? or ZeroMQ, TCP, and RAET?
15:23 Kelsar joined #salt
15:23 XenophonF I keep getting SaltReqTimeoutError using ZeroMQ and am transport-curious as a result.
15:23 XenophonF But from what I've read in the docs, it isn't clear how to make the switch or if I have to have a flag day.
15:24 XenophonF (a flag day would be bad b/c I'm already having minion connectivity problems and can't guarantee everyone will make the necessary config changes)
15:25 mchlumsky joined #salt
15:25 XenophonF cf. https://github.com/saltstack/salt/issues/26888
15:26 XenophonF the minion inits a new AsyncZeroMQReqChannel, then it inits a new AsyncAuth, and then it reports SaltReqTimeoutError
15:30 hammer065 joined #salt
15:36 racooper joined #salt
15:41 tiwula joined #salt
15:43 XenophonF I wonder if I should enable zmq_filtering.
15:52 babilen AFAIK RAET isn't really supported, so you might want to switch to the TCP transport right away (newer minions should try various)
15:58 saltslackbridge <gtmanfred> don’t use raet
15:59 saltslackbridge <gtmanfred> I know there was some work being done to try and make zmq and tcp able to be run at the same time
16:03 onlyanegg joined #salt
16:15 babilen gtmanfred: I remember some code/discussion about the minion trying both to enable users to switch from one to the other without losing all minions
16:15 saltslackbridge <gtmanfred> yup
16:15 saltslackbridge <gtmanfred> i remember it too, techhat was working on it.
16:15 babilen exactly
16:15 saltslackbridge <gtmanfred> but i don’t remember if it was actually finished
16:16 babilen I've never used it, but it was definitely "on the agenda" at one point
16:16 saltslackbridge <gtmanfred> lemme ask him
16:16 babilen ta
16:20 saltslackbridge <gtmanfred> apparently there were problems with the autodetect that were insurmountable at the time, so it was never finished
16:21 babilen That's a pity
16:21 saltslackbridge <gtmanfred> indeed
16:21 babilen Does that mean that they no longer are "insurmountable" ;) ?
16:23 girishb joined #salt
16:23 babilen So .. I guess we can't switch to TCP unless we roll out minion config changes via salt-ssh or ansible or something like that
16:23 saltslackbridge <gtmanfred> i do not know if it is still insurmountable. :slightly_smiling_face:
16:24 babilen I quite liked the idea of "try 0mq, nope .. try TCP, nope, ... try 0mq, nope, ... try TCP, WORKS!"
16:24 saltslackbridge <gtmanfred> agreed
16:24 XenophonF well shoot, that means flag day, doesn't it
16:25 XenophonF i found the master scaling doc and might play around with that first
16:26 XenophonF I have around 150 minions on this master so while changing the transport will be annoying, it won't be impossible
16:27 pipps joined #salt
16:27 babilen salt-ssh takes the minion cache as roster
16:27 babilen (which is nice as "oh, fuck" fallback)
16:29 XenophonF unfortunately for me, the master and its minions live on completely separate networks
16:29 XenophonF so salt-ssh isn't an option
16:29 babilen So no SSH between them?
16:29 XenophonF nope
16:29 babilen pity
16:29 XenophonF yeah
16:30 XenophonF i'd really like to figure out why connections timeout after new  AsyncAuth initialization
16:31 babilen The one time I was able to debug that we found that the MTU on one route was incorrectly set between two hops
16:32 babilen Debugging that one was a nightmare and there were many dead ends
16:32 XenophonF oh man I didn't think of an MTU problem but that could be it
16:32 XenophonF we've run into those before
16:32 XenophonF it sucks
16:32 babilen I was happy that it boiled down to a network issue eventually, but it was just "weird" in its behaviour
16:33 babilen Took me way too long to find too as I started to debug the "SaltStack angle" rather than the "communication angle"
16:34 XenophonF fiddling with the TCP keep-alive settings has helped a litte, but I'm still seeing those timeouts.
16:37 zer0def XenophonF: getting back to my ACM question, what would be the `acme` state module then? it does seem to be a bit lacking from the quick looks of things
16:38 zer0def we-e-e-ell, ok, it's oversimplified, that's for damn sure
16:39 XenophonF it's a thin wrapper around certbot that only does http-01
16:40 XenophonF i think
16:40 XenophonF which can be stateful, if the stars align
16:40 XenophonF it doesn't work at scale though
16:40 XenophonF e.g., i want to deploy a whole cluster of servers with LE certificates
16:41 zer0def well, it would require something more for dns-01, yeah
16:41 XenophonF I posted my planned approache for LE on salt-users@, which yes involved dns-01 validation
16:41 zer0def (i personally use dehydrated+dns-lexicon)
16:41 XenophonF for ACM you'd want to do the same thing: request the cert, validate using DNS not email
16:42 XenophonF problem is that validation takes a while
16:42 zer0def quite a significant while, in fact
16:42 babilen I use dehydrated too on some setups + reactors to make it usable behind LBs
16:42 XenophonF so really to do it properly you almost have to have orchestration or reactor
16:42 zer0def sounds like a reactor's job
16:42 XenophonF that's what i thought
16:43 sh123124213 joined #salt
16:44 zer0def anyone knows whether ACM is just ACME with additional parameters to make AWS happy?
16:44 XenophonF my idea was to run certbot + one of the DNS validation plugins on the master, fire an event locally via certbot's deploy-hook, and distribute the keymat to the appropriate minions via pillar
16:44 XenophonF as far as I can tell, ACM is their own thing
16:44 zer0def meh, i was hopeful
16:44 XenophonF you can only use it with amazon stuff
16:45 XenophonF I use it all over the place for CloudFront distributions or ELB/ELBv2 instances.
16:45 XenophonF but you can't give it a CSR and have it issue you a certificate
16:45 XenophonF and the only API is the AWS API
16:46 zer0def it's amazon, after all
16:46 XenophonF right
16:46 zer0def it would be far too liberal of them to allow usage of their certs outside of aws
16:46 XenophonF well, that's not the point of ACM
16:47 Bryson joined #salt
16:48 armyriad joined #salt
16:48 shiranaihito joined #salt
16:51 deuscapturus joined #salt
17:06 deuscapturus_ joined #salt
17:08 deuscapturus_ joined #salt
17:12 Deliant joined #salt
17:18 deuscapturus joined #salt
17:20 Deliant joined #salt
17:24 aruns__ joined #salt
17:25 pipps joined #salt
17:30 Deliant joined #salt
17:33 deuscapturus left #salt
17:34 jeffspeff can you use an onlyif on a cmd.run state?
17:34 oida joined #salt
17:35 tflo joined #salt
17:35 saltslackbridge <gtmanfred> you can use onlyif in any state
17:35 saltslackbridge <gtmanfred> https://docs.saltstack.com/en/latest/ref/states/requisites.html#onlyif
17:37 jeffspeff gtmanfred, here's the state https://gist.github.com/jeffclay/400118c9130fd52238f18a185c413d43 if I run the "powershell Test-Path -Path "C:\ProgramData\chocolatey\lib\flashplayerplugin" -PathType Container" from cmd it returns false. however, in salt it runs the rmdir command anyways.
17:39 saltslackbridge <gtmanfred> how does windows handle `cmd.retcode`?  what do you get back when you run `salt-call cmd.retcode 'Test-Path Path "C:\ProgramData\chocolatey\lib\flashplayerplugin" -PathType Container' shell=powershell`
17:39 pipps joined #salt
17:41 jeffspeff just added the results to the comments on that gist
17:42 saltslackbridge <gtmanfred> well that is fun.
17:42 tflo joined #salt
17:44 saltslackbridge <gtmanfred> flip the quotes maybe? i don’t use windows often.
17:44 saltslackbridge <gtmanfred> `salt-call cmd.retcode "Test-Path Path 'C:\ProgramData\chocolatey\lib\flashplayerplugin' -PathType Container" shell=powershell`
17:45 jdipierro joined #salt
17:46 jeffspeff i updated the the comment
17:47 saltslackbridge <gtmanfred> what is the output when you actually run the state in the first comment?
17:48 tflo joined #salt
17:49 jeffspeff updated gist
17:49 relidy joined #salt
17:50 saltslackbridge <gtmanfred> yeah, so retcode doesn’t work on windows the same way it does on linux, because if that file doesn’
17:50 saltslackbridge <gtmanfred> t exist on linux, it would return a 1, (test -f /path/to/file)
17:51 saltslackbridge <gtmanfred> i don’t know how to fix that on windows
17:51 saltslackbridge <gtmanfred> might just have to go back to wrapping the states in jinja
17:51 jeffspeff ok, give me a minute to revert the states and test that again.
17:54 gh34 joined #salt
17:54 zer0def gtmanfred: just to follow up, onlyif applies to all states, like unless, yes?
17:55 zer0def actually, that's in state requisites, so it should
17:57 saltslackbridge <gtmanfred> yes
17:58 onlyanegg joined #salt
18:04 mikecmpbll joined #salt
18:09 edrocks joined #salt
18:20 pipps joined #salt
18:30 xet7 joined #salt
18:34 Goobhala joined #salt
18:37 Goobhala left #salt
18:38 Aleks3Y joined #salt
18:42 demize joined #salt
18:53 pipps joined #salt
19:02 XenophonF apparently I can start both the TCP and 0mq transports on my master after all: https://docs.saltstack.com/en/latest/ref/configuration/master.html#transport-opts
19:04 vinnu joined #salt
19:08 pipps joined #salt
19:10 pualj joined #salt
19:16 mikecmpbll joined #salt
19:19 saltslackbridge <gtmanfred> yes, you just can’t have the minion fail over between the two
19:22 tikedbus joined #salt
19:29 XenophonF i love render errors in "<string>" );
19:32 pipps joined #salt
19:35 Trauma_ joined #salt
19:40 pipps joined #salt
19:47 pualj_ joined #salt
19:50 Hybrid joined #salt
19:51 sh123124213 joined #salt
19:51 Edgan XenophonF: Do you use pillar encryption in some form?
19:59 systemexit joined #salt
20:04 onlyanegg joined #salt
20:04 pualj joined #salt
20:07 pipps joined #salt
20:12 pualj_ joined #salt
20:15 edrocks joined #salt
20:20 onlyanegg joined #salt
20:22 pipps joined #salt
20:33 pualj joined #salt
20:45 sjorge joined #salt
20:52 pualj_ joined #salt
21:05 XenophonF yes
21:09 onmeac joined #salt
21:11 nledez joined #salt
21:20 Hybrid joined #salt
21:22 XenophonF Edgan: this is how I do GPG-encrypted Pillars - https://www.reddit.com/r/saltstack/comments/6py6a0/emacs_epafile_import_text_gpg_renderer/
21:22 XenophonF for example - https://github.com/irtnog/salt-pillar-example/tree/master/login/example/com
21:29 XenophonF I would love to know why this JSON piped to |yaml was causing a render error
21:29 XenophonF http://ix.io/DC7
21:34 pipps joined #salt
21:36 pipps99 joined #salt
21:37 XenophonF this is the error generated
21:37 XenophonF http://ix.io/DCb
21:37 XenophonF I'm importing the JSON file using import_json
21:38 XenophonF and other files get imported no problem
21:38 pipps_ joined #salt
21:39 pualj joined #salt
21:41 pipps joined #salt
21:45 DammitJim joined #salt
21:47 pipps joined #salt
21:47 Edgan XenophonF: How is performance? I am using another method that is horrible performance.
21:48 XenophonF I haven't noticed any problems but I'm doing using it at scale.
21:48 Edgan XenophonF: not using?
21:51 Edgan XenophonF: how do you define at scale?
21:53 XenophonF more than 100 minions
21:54 XenophonF I'm using it in a Salt deployment that has around 50 minions.
21:57 pipps joined #salt
22:00 Edgan XenophonF: I am thinking of going with the model of secure.sls, and letting it encrypt all sensitive data in one pass. I suspect overhead of a decrypt call will always be greater than the variable amount of data.
22:15 Bryson joined #salt
22:18 pipps99 joined #salt
22:20 MTecknology 100 minions isn't really hitting scale
22:20 MTecknology that's still hobby time :)
22:28 ipmb joined #salt
22:29 XenophonF i know it :)
22:29 XenophonF the contract where i have 150 minions, they're all located in far-away places
22:30 XenophonF which complicates things in interesting ways
22:30 MTecknology For my salt presentation, I want to spin up 100 minions on a local cluster and show them being auto-deployed/provisioned/added to load balancer/added to dns/etc.
22:30 XenophonF dammit - so everywhere I use {% import_json ... as foo %} and {{ foo|yaml }} has broken
22:31 MTecknology My previous employer had me dealing with ~200 remote facilities, all connected via T1's.
22:31 XenophonF I just upgraded from 2017.7.1 to 2017.7.2 - http://ix.io/DCk
22:32 XenophonF MTecknology: I have about 50% server coverage with Salt, and most of that's in AWS
22:33 XenophonF maybe Jinja is too new?  I don't remember what version I was running before
22:33 XenophonF this is on FreeBSD
22:33 MTecknology what do you actually have that's breaking?
22:35 XenophonF I'm importing JSON like this - http://ix.io/DC7 - and rendering it in my Pillar SLS using {{ var1|yaml }}, but since upgrading Salt that's throwing errors like this - http://ix.io/DCb
22:38 MTecknology what's in the pillar?
22:39 wongster80 joined #salt
22:40 XenophonF hang on i'll post it on gist
22:42 XenophonF https://gist.github.com/xenophonf/2eb169eae43ab74844973381ad0e33f7
22:44 pipps joined #salt
22:45 DoomPatrol dang that is one heck of a pillar
22:45 pipps99 joined #salt
22:46 MTecknology holy frick
22:46 XenophonF well, it's my production salt master config
22:49 XenophonF I like having the AWS stuff in JSON because that looks more native.
22:50 XenophonF I guess I can inline it all
22:54 sjorge joined #salt
22:55 cyteen joined #salt
23:01 MTecknology could you write a module to safely read values and return verbose errors?
23:08 pipps joined #salt
23:09 pipps99 joined #salt
23:15 sjorge joined #salt
23:16 masber joined #salt
23:18 cholcombe joined #salt
23:27 oida joined #salt
23:29 wongster80 joined #salt
23:30 kettlewell joined #salt
23:40 notCalle joined #salt
23:40 onlyanegg joined #salt
23:42 pipps joined #salt
23:48 shadoxx Not sure who was in here the other day, but I was able to figure out my issue with startup_states in the salt-cloud map file
23:48 shadoxx i was referencing my states with a '/' in the name instead of a '.'
23:48 shadoxx now everything seems to be working as expected
23:51 oida joined #salt

| Channels | #salt index | Today | | Search | Google Search | Plain-Text | summary