Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-06-17

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

All times shown according to UTC.

Time Nick Message
00:06 emaninpa joined #salt
00:13 hal58th__ joined #salt
00:17 hal58th joined #salt
00:20 baweaver joined #salt
00:23 otter768 joined #salt
00:34 aqua^c joined #salt
00:36 mrbigglesworth joined #salt
00:37 nyx_ joined #salt
00:44 rdavis Guessing pull requests are checked like throughout the day/night.
00:45 Micromus joined #salt
00:48 baweaver joined #salt
00:49 bones050 joined #salt
00:52 moos3 joined #salt
00:54 FeatherKing joined #salt
01:02 cheine_ joined #salt
01:03 aqua^c joined #salt
01:05 Vynce joined #salt
01:08 baweaver joined #salt
01:12 hrumph joined #salt
01:12 hrumph yo
01:12 hrumph i'm thinking of installing salt a work so i can communicate from a linux machine to windows pc's through powershell at the very least and then maybe start using the other capabilities of salt over time
01:13 hrumph can salt get a list of all installed software by itself or does it only think in terms of stuff installed with a salt repo?  that's one question
01:13 Vynce if a master has many IP interfaces, do i have to do anything special to make sure it's listening as a salt master on all of them?
01:14 Gilou Vynce, default is 0.0.0.0
01:14 hrumph the other question is if i have user machines such that i deploy the software with salt to the extent possible.....how do windows updates factor into this?
01:14 Vynce (i'm tryign to debug why a new minion isn't showing up in the unauthorized keys list, that jsut seems like an obvious place to start)
01:14 Vynce Gilou: … which means, no?  I'm not sure what ramifications that has, since afaik 0.0.0.0 isn't a real IP address...
01:16 Gilou it means yes
01:16 Gilou 0.0.0.0 is a way to say all interfaces :)
01:16 Vynce OK. … well, so is ther ea debug checklist for minion-doesn't connect-to-master?
01:16 Vynce (and thanks)
01:16 Gilou firewall?
01:17 Vynce minion can ping master at the IP /etc/salt/minion specifies
01:17 Gilou try hitting port 4506 of the master from the minion ?
01:18 desposo joined #salt
01:18 hrumph how many people actually control windows machines with salt?
01:18 Gilou hrumph, it works nicely, depending on how far you want to go
01:18 Gilou I had no expectations, but the chocolatey module is really nice to have in a mixed environment ;)
01:18 hrumph how does it differ from the sort of stuff you can do with active directory (forgetting about authentication)
01:18 Gilou it does not, but it does not depend on it either..
01:18 Vynce Gilou: any particular method recommended?  telnet <IP> 4506 says "Trying <IP>…"
01:19 Gilou Vye, not good, 4505 ?
01:19 Gilou netcat -v or telnet should work, and be able to initiate a TCP connection
01:20 Vynce 4505 also just says Trying… so i guess i look at firewall next.  thanks.
01:20 Gilou cf http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
01:20 Vynce does it jsut need those two ports? … ah, thanks.
01:20 Gilou (yeah)
01:21 Vynce hm.  ufw, but could not find profiel salt.
01:21 Gilou hmm, the package should have installed it
01:22 Gilou maybe it only does so on ubuntu
01:22 Gilou or maybe only on master
01:22 Gilou dunno, anyway: https://github.com/saltstack/salt/blob/develop/pkg/salt.ufw
01:22 Gilou :)
01:23 Vynce that files says to run ufw app update, but (my) ufw does not recognize that as a valid command
01:24 Gilou I have no experience with ufw
01:24 Gilou but I guess there is something approaching..
01:25 david_an111 joined #salt
01:26 Gilou (but if you don't know ferm, it's a nice tool :D)
01:27 beauby joined #salt
01:27 Vynce never heard of it. (<<— not a sysad)
01:28 Gilou though here you are, playing with salt ;))
01:28 Gilou ferm is a tool to configure in a saner way iptables rules, quite devops friendly
01:30 __number5__ Gilou: does ferm understand docker networking? those stuff made me crazy
01:31 Vynce Gilou: sadly, yes.  but thanks; i jsut added the ruels to ufw manually.  seems to've worked!  another day, another problem created and solved. ( :
01:32 Gilou I don't use docker (yet), but well.. if you can do it using iptables, then, ferm probably can handle it yeah
01:34 Gilou __number5__, quick google lookup gives https://pypi.python.org/pypi/Ferment/0.0.1 .. doesn't seem like show time ready ;)
01:34 rdavis_ joined #salt
01:34 Gilou Vynce, heh... too much security & all that ;))
01:35 Gilou well at least now it can talk to its master, and you can tell ufw to go f*ck itself in a state.. :)
01:36 ageorgop joined #salt
01:39 c10 joined #salt
01:40 aqua^c joined #salt
01:45 fllr joined #salt
01:45 ageorgop joined #salt
01:47 ilbot3 joined #salt
01:47 Topic for #salt is now Welcome to #salt | 2015.5.1 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:49 Vynce left #salt
02:04 beauby joined #salt
02:06 MatthewsFace joined #salt
02:12 aqua^c joined #salt
02:15 __number5__ Gilou: ferment looks promising, thanks :)
02:17 aparsons joined #salt
02:19 aparsons_ joined #salt
02:28 beauby joined #salt
02:31 AlexStraunoff joined #salt
02:32 moos3 joined #salt
02:32 writtenoff joined #salt
02:33 darknight87 joined #salt
02:39 aqua^c joined #salt
02:41 berserkir joined #salt
02:44 favadi joined #salt
02:47 ageorgop joined #salt
02:49 donmichelangelo joined #salt
02:54 onorua joined #salt
02:55 otter768 joined #salt
03:01 rdavis Anyone know what's good to search for in the jenkins output to find the cause of a test failure?
03:02 eth2 left #salt
03:02 rbatiz joined #salt
03:03 rdavis Nevermind, I think I found it.
03:04 rdavis Nope, might not be it, nevermind.
03:08 rdavis left #salt
03:08 svetlana left #salt
03:10 beauby joined #salt
03:20 Bryson joined #salt
03:21 hrumph if i set everything up with salt, and supposing that wake up on lan doesn't work most of the time and sometimes a lot of the user machines are off (something i can't control) then what happens if i push out updates?
03:25 aqua^c joined #salt
03:25 cheus joined #salt
03:27 hrumph how does salt deal with pushing updates to minions that aren't on?
03:27 c10 joined #salt
03:31 sgargan joined #salt
03:38 Sage|4U_ joined #salt
03:42 Sage|4U_ left #salt
03:44 mixja joined #salt
03:59 hrumph what would happen if salt pushed an update and somneone turned of their machine durning the update causing registry corruption? what could prevent that?
04:00 hrumph does it make any sense to try to use salt for windows desktop provisioning?
04:01 ajw0100 joined #salt
04:11 loz-- joined #salt
04:12 Furao joined #salt
04:13 aridere joined #salt
04:25 hasues joined #salt
04:25 hasues left #salt
04:28 cilkay hrumph: That's not an issue that is specific just to SaltStack. It could happen with any deployment or CM system. I think the philosophy of SS is that you can always describe the state in which your machine needs to be and get the machine back in that state.
04:29 cilkay That means you'll have to treat the workstations as ephemeral, which you're probably doing if you have something you'd like to automate with SS anyway.
04:30 cilkay In other words, you should be able to nuke and pave those Windows boxes without any side effects, like losing data, because any data would be on a file server or in the cloud somewhere.
04:30 hrumph cilkay, i'd like to automate, but i need be sure that other poeple in the same sitation would also do it so i'm not doing something crazy
04:31 p0rkbelly joined #salt
04:31 cilkay I haven't used it for automating Windows but I don't see why it's any different than automating anything else. Dirty shutdowns can cause problems for other operating systems, too.
04:32 hrumph cilkay, if you've got a bunch of servers you can basically guarantee that they are on all of the time
04:32 cilkay In other words, I don't think this is a SaltStack issue as much as a policy issue for your organization.
04:32 cilkay Indeed.
04:32 cilkay If the policy is "no data that you care about stored on workstations", who cares if the workstations break? You can always provision them and bring them back to a known good state.
04:33 cilkay I'm not sure that you need "a bunch of servers". Modern Windows servers with HyperV can do some pretty impressive things and I'm not a Windows guys.
04:34 hrumph cilkay, right let's just say that i had perfect control then theses workstations would be on all of the time and no one would be able to turn them off even if this isn't really a good idea from the point of view of power conservation
04:35 hrumph so my management software could do whatever was needed at any time without the posibility of interference
04:35 cilkay You could do Wake on LAN and provision when they're not being used and then turn them off.
04:35 cilkay I've done that sort of thing with the open source project "ntinstall" in the past.
04:35 hrumph cilkay, i'm almost certain that wol doeosn't work on my organisations lan but i'll test that one more time
04:37 lightslategray joined #salt
04:37 hrumph how does salt deal with windows updates?
04:38 ITChap joined #salt
04:38 cilkay I don't know. I only started using SS last week and have been using it to provision a Vagrant onto which I install and configure a Node.js and AngularJS application.
04:39 cilkay a Vagrant VM, that is
04:39 hrumph it's going to be complicated..the windows updates you need depends so much on what you've installed....
04:40 loz-- joined #salt
04:40 cilkay I can go from the Debian netinstall image from Debian.org to a working server in under an hour all provisioned automatically.
04:42 agj joined #salt
04:42 cilkay That is probably a job for SS pillars.
04:45 hrumph would be nice if windows one-get existed for windows 7, then saltstatck could presumably plug into it
04:46 lightslategray left #salt
04:50 ramteid joined #salt
04:55 aarontc_ joined #salt
04:56 otter768 joined #salt
04:58 c10 joined #salt
05:00 joeto joined #salt
05:06 pppingme joined #salt
05:17 catpiggest joined #salt
05:22 camelType joined #salt
05:29 keimlink joined #salt
05:31 freelock joined #salt
05:44 evle joined #salt
05:51 Xiao joined #salt
05:52 freelock joined #salt
05:55 favadi joined #salt
06:00 colttt joined #salt
06:01 AndreasLutro joined #salt
06:06 stoogenmeyer joined #salt
06:08 dgk joined #salt
06:08 colttt joined #salt
06:13 MatthewsFace joined #salt
06:13 mixja joined #salt
06:17 prathee joined #salt
06:21 mrbigglesworth joined #salt
06:24 Furao joined #salt
06:28 flyboy joined #salt
06:28 prathee joined #salt
06:33 AndreasLutro anyone know if it's possible to have salt-cloud profile "defaults", or for one profile to extend another?
06:33 AndreasLutro just as I write that I find it in the documentation!
06:34 aarontc joined #salt
06:35 KermitTheFragger joined #salt
06:37 julez joined #salt
06:40 trikke joined #salt
06:44 c10 joined #salt
06:49 ndrei joined #salt
06:56 otter768 joined #salt
06:57 ggoZ joined #salt
07:02 kawa2014 joined #salt
07:05 eseyman joined #salt
07:05 joeto joined #salt
07:07 mrbigglesworth joined #salt
07:08 onorua joined #salt
07:15 Romlok joined #salt
07:17 Furao AndreasLutro: yes it’s documented in salt-cloud doc
07:17 Furao somekind of extend
07:17 c10 joined #salt
07:21 lb1a joined #salt
07:22 favadi joined #salt
07:28 supersheep joined #salt
07:34 markm joined #salt
07:35 MatthewsFace joined #salt
07:37 bitmand joined #salt
07:38 Grokzen joined #salt
07:47 linjan joined #salt
07:54 fllr joined #salt
07:54 moos3 joined #salt
07:54 mrbigglesworth joined #salt
07:58 impi joined #salt
08:04 s_kunk joined #salt
08:05 s_kunk joined #salt
08:07 Xevian joined #salt
08:10 chiui joined #salt
08:13 mrbigglesworth joined #salt
08:17 moos3 joined #salt
08:18 al joined #salt
08:19 toanju joined #salt
08:21 lothiraldan joined #salt
08:21 kawa2014 joined #salt
08:26 julez joined #salt
08:28 aqua^c_ joined #salt
08:28 clmsy joined #salt
08:29 kawa2014 joined #salt
08:32 clmsy left #salt
08:32 fllr joined #salt
08:33 julez joined #salt
08:33 fllr joined #salt
08:37 AndreasLutro I'd like to have a separate top.sls or similar that's put in .gitignore for a local test vm but can't seem to figure out a way to do it with environments - has anyone done something like this?
08:49 kawa2014 joined #salt
08:52 magicmonkey joined #salt
08:53 illern joined #salt
08:53 Arendtsen joined #salt
08:53 epcim joined #salt
08:54 AndreasLutro nevermind, was just an issue of re-ordering the file and pillar roots
08:54 magicmonkey hi everyone - I wondered if anyone could help me with a salt mine problem I'm having?  It's probably quite easy, but I've really only just started using salt mine (although I've used salt states for a while)
08:55 magicmonkey I'm trying to use a list of IP addresses for minions which match a particular grain in a config file (typical reverse-proxy load-balancing a set of webservers)
08:55 magicmonkey however
08:55 magicmonkey when I run the state.highstate command on the proxy minion, then instead of getting a list of the currently-matching minions, I get a list of *all* minions which have ever matched in the past
08:56 magicmonkey I've tried every type of cache clear I can find, but I still get a huge list of none-existant servers (which previously existed, but have since been shut down)
08:56 magicmonkey anyone come across this in the past / can point me in the right direction to get a currently-accurate list of minions in my config file?
08:57 magicmonkey as an example, if I run "salt-call mine.get 'ec2_roles:webserver' network.ip_addrs grain" on a proxy minion, then I get the huge list instead of the currently-correct list
08:57 babilen magicmonkey: So, you *did* run "salt '*' mine.flush" already? (or what does "every type of cache clear I can find" mean exactly?)
08:57 otter768 joined #salt
08:58 magicmonkey babilen: Yes, I think so, but let me run it again just in case, 2 secs...
08:58 babilen "salt-run cache.clear_mine" too
08:58 babilen "salt-run cache.clear_all" even if you want to try that
08:59 * babilen would expect "salt-run cache.clear_mine" to sort you out
09:00 magicmonkey (2 more secs, just grappling with another problem - I'm running this on Ubuntu and I just updated all of the salt packages in case it was a bug which had been fixed, but I'm getting a different error...  just fixing that...)
09:00 chiui joined #salt
09:00 babilen no, do it now! ;)
09:01 magicmonkey heh :)
09:01 magicmonkey restarting all of the salt minions fixed my other problem (a msg format incompatibility or something like that)
09:01 magicmonkey 1) after "salt '*' mine.flush" I still get the big list
09:02 magicmonkey 2) after "salt-run cache.clear_mine" (which returns "False") I still get the big list
09:02 magicmonkey 3) after "salt-run cache.clear_all" (which returns "False") I still get the big list
09:02 magicmonkey :-/
09:02 julez joined #salt
09:02 babilen Anything interesting in the master log when you run either
09:02 babilen "salt-run cache.clear_mine" or "salt-run cache.clear_all" ?
09:02 magicmonkey I've currently only got 1 webserver running, but I get a list of 30-odd
09:03 magicmonkey babilen: checking...
09:03 magicmonkey oh, if I run it with "-l debug" then I see this:
09:03 magicmonkey [DEBUG   ] Error loading runners.nacl: libnacl import error, perhaps missing python libnacl package
09:03 magicmonkey 2015-06-17 09:03:29,680 [salt.loader      ][DEBUG   ][6041] Error loading runners.nacl: libnacl import error, perhaps missing python libnacl package
09:04 babilen https://github.com/saltstack/salt/issues/21986 if probably worth reading
09:04 magicmonkey checking, certainly sounds relevant...
09:04 fredvd joined #salt
09:04 ingslovak joined #salt
09:05 babilen Try: salt-run cache.clear_all tgt='*'
09:05 magicmonkey trying that now
09:05 babilen sorry, I forgot that you have to provide the tgt expression there
09:06 babilen (it should now return True)
09:06 magicmonkey that returned "True", however I still get the big list of servers and I have lots of dead minions referenced in /var/cache/salt/master/minions/
09:07 magicmonkey I could just empty that dir I guess
09:08 magicmonkey but I was hoping there was a handy salt command I could run, it feels less like I'm interfering with the internal functionality
09:08 magicmonkey (thanks for helping, by the way!)
09:10 babilen Looks as if that is a genuine bug and as if you have to clear the contents of /var/cache/salt/minions/<dead_minions>
09:11 magicmonkey I've added a comment to that bug you linked to.  I'll clear that dir for now, and keep an eye on that bug report.  Thanks again for your help, I was going a little bit crazy not being able to figure it out!
09:14 babilen You are welcome
09:15 impi joined #salt
09:15 babilen (don't forget to update the mine with the current information)
09:22 sgargan joined #salt
09:23 magicmonkey ah yes, well pointed out :)  It now seems to be working, although I got my minions in a bit of a state and just ended up rebooting everything
09:23 magicmonkey after that, emptying that dir and running mine.update everywhere has helped
09:23 magicmonkey ta!
09:29 ITChap joined #salt
09:29 babilen \o/
09:30 mrbigglesworth joined #salt
09:32 supersheep joined #salt
09:33 briner joined #salt
09:34 moos3 joined #salt
09:40 julez joined #salt
09:44 c10 joined #salt
09:46 ndrei joined #salt
09:48 _ektat_ joined #salt
09:48 _ektat_ hi there
09:51 Micromus joined #salt
09:51 ange hello
09:51 ange a simple way to check if a url return a 200 from a state ?
09:52 _ektat_ joined #salt
09:53 ange http://docs.saltstack.com/en/latest/ref/states/all/salt.states.http.html apparently
09:54 _ektat_ is it possible to run a master with multiple syndic controlling only salt-ssh  (box with no minion daemon)? I guess it's pretty good for running salt on a lossy mesh network
09:56 ndrei joined #salt
09:59 salt_newbie joined #salt
09:59 _mel_ joined #salt
10:00 sgargan joined #salt
10:00 salt_newbie I change salt-master's public_port, and then salt won't work. How do I config salt-minion for this issue?
10:01 salt_newbie I don't found any thing about public port in minion config.
10:01 keimlink joined #salt
10:02 mou joined #salt
10:02 AndreasLutro salt_newbie: http://docs.saltstack.com/en/latest/ref/configuration/minion.html#master-port ?
10:02 darknight87 joined #salt
10:03 AndreasLutro mm no, that's for ret_port
10:04 unused_phd joined #salt
10:05 salt_newbie in master side | in minion side
10:05 salt_newbie ret_port  == master_port
10:05 linjan joined #salt
10:06 salt_newbie that's ok.
10:07 salt_newbie i mean public_port: 4505 can't change.
10:07 AndreasLutro looks like it from the docs. I'll have a quick look in the code
10:09 _ektat_ salt-syndic with salt-ssh "minions without minion daemon" , is it do-able ?
10:10 AndreasLutro salt_newbie: it looks like the minion is supposed to receive the publish port from the master automatically, you shouldn't need to configure the minion
10:11 AndreasLutro of course it is possible that there's a bug but that's what the code looks like. you could open a github issue if it doesn't seem to work
10:13 salt_newbie public_port  ==  ?
10:16 joeto1 joined #salt
10:16 cwyse joined #salt
10:21 flebel joined #salt
10:22 fllr joined #salt
10:23 ndrei joined #salt
10:25 desposo joined #salt
10:26 bluenemo joined #salt
10:29 slav0nic joined #salt
10:36 unusedPhD joined #salt
10:36 unused_phd joined #salt
10:41 Arendtsen If I run salt '<hostname>' state.sls users and I get an Invalid argument. What am I missing?
10:41 nyx__ joined #salt
10:41 AndreasLutro Arendtsen: you need to specify which sls to run
10:41 AndreasLutro state.sls my_statefile
10:42 ronrib joined #salt
10:42 Arendtsen AndreasLutro: I have the file users created
10:44 sgargan joined #salt
10:47 giantlock joined #salt
10:50 bitmand left #salt
10:52 ndrei joined #salt
10:54 supersheep joined #salt
10:58 otter768 joined #salt
11:01 ndrei joined #salt
11:04 TyrfingMjolnir joined #salt
11:04 keimlink joined #salt
11:08 lobi joined #salt
11:16 darknight87 joined #salt
11:17 Hazelesque Hi there
11:17 ndrei joined #salt
11:17 Arendtsen Is this expected:
11:17 Arendtsen root@cm01:/usr/local/etc/salt/pillar/users # salt 'cm01.arendtsen' test.ping
11:17 Arendtsen Invalid argument
11:17 Hazelesque if I wanted to be able to, say, audit the cloud servers (VMs) that exist against the hosts that are known to salt (i.e. have a salt-key accepted)...
11:17 Hazelesque can I use the salt-cloud stuff just "read only"?
11:18 Hazelesque i.e. can I set up a Provider, but with no Profile details
11:18 Hazelesque and just use it to pull down info about our cloud environment?
11:19 Hazelesque e.g. to generate a list of hosts that still need salt bootstrapping?
11:19 Hazelesque (I mean, I could just use the nova CLI and mangle the output, but we have ~150 VMs at any one time and it's a moving target... and doing this repeatedly is getting tedious...)
11:20 Hazelesque and if Salt can expose that information usefully (e.g. with the "cloud.query" execution module?), then it saves me writing it against something provider-specific, or against libcloud directly
11:20 Hazelesque is it "safe" to do this?
11:24 * Hazelesque has a thought...
11:24 * Hazelesque goes and creates a read-only API key
11:24 OliverMT anyone tried using saltstack to provision circleci build stacks?
11:24 ggoZ joined #salt
11:28 epcim joined #salt
11:32 samnmax joined #salt
11:32 supersheep joined #salt
11:34 giantlock joined #salt
11:55 dunz0r Can I do a regexp match on grains in my pillar?
11:55 dunz0r Like {% elif grains['os'] == '(CentOS|RedHat)' %}   ?
11:57 onorua joined #salt
11:58 hunmaat dunz0r: os_family is for you
11:58 dunz0r Oooooh
11:58 dunz0r hunmaat: Didn't know about that one
11:59 darknight87 joined #salt
11:59 dunz0r Is debian and ubuntu considered to be the same family?
11:59 lothiraldan joined #salt
12:00 dunz0r They are... hmm
12:02 hunmaat but if they weren't, you could still use `grains['os'] in ["Debian", "Ubuntu"]` instead of a regexp
12:04 Zytox joined #salt
12:04 ndrei joined #salt
12:10 mapu joined #salt
12:11 echoes86 joined #salt
12:11 fllr joined #salt
12:11 echoes86 Hello ... I'm from Italy and I've a problem with salt configuration when Master and Minion are of different versions
12:12 echoes86 Have you ever had to resolve this kind of problem?
12:12 TyrfingMjolnir joined #salt
12:23 pacopablo joined #salt
12:24 ggoZ joined #salt
12:24 KennethWilke joined #salt
12:25 tmclaugh[work] joined #salt
12:26 babilen echoes86: Your master has to be at least as new as your minions and you should try to ensure that they are the same version
12:26 bluenemo joined #salt
12:26 babilen Which versions do you have on the master and minions?
12:27 pickledpig joined #salt
12:27 julez joined #salt
12:42 echoes86 Thank you for your replies
12:42 echoes86 Master: 0.17.5, Minion: 2015.5.0
12:43 echoes86 Master is on Ubuntu 14.04, Minion on Ubuntu 12.02
12:43 echoes86 So I cannot upgrade my minion to the last version
12:43 echoes86 Can you tell me how I can do?
12:45 timoguin joined #salt
12:46 PI-Lloyd minion is already newer than the master. Just update salt-master and you should be good to go
12:46 iwishiwerearobot joined #salt
12:49 echoes86 salt-master is already the newest version
12:49 echoes86 This is the output of my command: apt-get install --only-upgrade salt-master
12:50 echoes86 Thanks for your reply @PI-Lloyd
12:51 AndreasLutro echoes86: make sure the salt PPA is added on the master, as described here http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
12:52 rawzone joined #salt
12:53 iwishiwerearobot joined #salt
12:54 subsignal joined #salt
12:54 babilen echoes86: You should upgrade your master to 2015.5.0 -- http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
12:54 babilen ah
12:55 echoes86 You're my savers ... thank you very very very much
12:55 echoes86 And sorry for my stupid question :)
12:55 babilen It's not stupid at all and, in fact, a common problem
12:55 echoes86 Thank you again :)
12:56 babilen The problem is that salt is both packaged in Debian (and thereby Ubuntu, ...) and repositories maintained by saltstack itself.
12:57 echoes86 Yes ... I forgot to execute the "apt-get update" after having addedd the ppa
12:59 otter768 joined #salt
13:01 chiui joined #salt
13:02 lobi Hello @all -  What's the difference of the main purpose between state modules and execution modules? I ask because I want to define a btrfs state – and found that the btrfs module is an execution module so I only can use it with module.run. Why is there no btrfs state module – I thought it makes more sense – but perhaps I misanderstood something.
13:07 amcorreia joined #salt
13:07 AndreasLutro lobi: modules are like functions. states can build on top of modules and have a strict return value format that can be used by the state system. there's no big underlying difference between the two
13:09 dyasny joined #salt
13:11 FeatherKing joined #salt
13:11 dRiN joined #salt
13:14 emaninpa joined #salt
13:15 kulty joined #salt
13:16 aparsons joined #salt
13:18 bhosmer joined #salt
13:23 hasues joined #salt
13:23 hasues left #salt
13:25 joshprewitt joined #salt
13:26 lobi AndreasLutro: OK, but if I run a state.highstate I get the status of the states I defined , but the result of module.run seems always to be true. So if I want to define a btrfs state which “returns” the same type o output like all the other states created with state modules I have to write mx own btrfs state module?
13:28 darknight87_ joined #salt
13:30 AndreasLutro lobi: what module function are you calling that's only returning True?
13:30 darknight87 joined #salt
13:31 racooper joined #salt
13:35 rojem joined #salt
13:36 darknight87 joined #salt
13:38 ndrei joined #salt
13:40 DammitJim joined #salt
13:42 anotherZero joined #salt
13:42 jdesilet joined #salt
13:42 ctolsen joined #salt
13:42 _JZ_ joined #salt
13:43 kaptk2 joined #salt
13:44 debian112 joined #salt
13:47 murrdoc joined #salt
13:48 mpanetta joined #salt
13:50 Ahlee It's expected behavior that if you call state.highstate that requires saltutil.sync_all items (_modules, _states, _grains, etc) that those newly sync'd items aren't available to the existing highstate
13:52 murrdoc its annoying
13:52 murrdoc but its known
13:52 murrdoc Ahlee:  do u use reactors ?
13:53 Ahlee murrdoc: yes
13:53 Ahlee hrm.
13:53 Ahlee i wonder if my reactor is syncing just one thing
13:53 murrdoc this look ok to u  ? https://gist.github.com/anonymous/8d174088f25451c55b9a
13:54 ekristen joined #salt
13:55 Ahlee i see nothing wrong with it
13:55 Ahlee should sync_all, refresh-pillars, run state.sls global.minion, then highstate
13:55 numkem joined #salt
13:55 Ahlee actually really clever. I'm totally stealing that
13:55 murrdoc yay
13:56 murrdoc your approval feels good man
13:56 murrdoc how was the meetup
13:56 murrdoc i was in la from work
13:56 Ahlee it was ok
13:56 Ahlee i won't be hosting any more with ITA
13:56 murrdoc why what they do
13:57 Ahlee i was stronga rmed into turning the meetup into a recruiting event for them
13:57 Ahlee oh well
13:57 murrdoc not professional
13:57 murrdoc chicago is a tough market i guess
13:57 Ahlee Not at all.
13:58 Ahlee but, the damage has been done
13:58 Ahlee I really appreciate those that came and refused to give their email
13:58 kulty joined #salt
13:59 soren joined #salt
13:59 murrdoc decent turnout ?
14:00 favadi joined #salt
14:00 Ahlee 20ish?
14:00 fllr joined #salt
14:00 Ahlee I love having discussions with myself on closed PRs in our stash instance
14:00 Ahlee why can't we hire people that care?
14:01 ange any idea about how to do a healthcheck on a http server from the state ? tryed with - name: '[ `curl -s -o /dev/null -I -w "%{http_code}"  http://{{ grains["fqdn_ip4"][0] }}/about` == "200" ]' not really working it seems
14:03 murrdoc use module.run and cp.get_url
14:07 colttt joined #salt
14:08 TyrfingMjolnir joined #salt
14:09 undefinednull joined #salt
14:10 andrew_v joined #salt
14:11 CeBe joined #salt
14:13 toanju joined #salt
14:14 ange sounds odd
14:14 kbyrne joined #salt
14:15 zerthimon joined #salt
14:16 zerthimon Why does salt-minion process take 100% cpu ?
14:16 markm_ joined #salt
14:19 coval3nce joined #salt
14:19 Ahlee is it processing a large state file?
14:20 Ahlee logging enabled? showing anything?
14:20 zerthimon Ahlee: it's just sitting there reading /proc/686/net/tcp
14:20 Ahlee i've seen it 100% when there's a corrupt cache value, and the minion's actually stuck in a restart loop (windows host in this case)
14:20 Ahlee zerthimon: interesting, haven't seen that one. Get that from strace?
14:20 racooper joined #salt
14:21 zerthimon Ahlee: yeah
14:21 zerthimon Ahlee: then a process dies and new one starts doing the same thing
14:22 Ahlee zerthimon: any file.managed off the salt filesystem?
14:22 Ahlee er, s/filesystem/file server/
14:22 zerthimon Ahlee: nope
14:22 Ahlee can you stop salt-minion and kill any processes left over?
14:22 Ahlee maybe it's trying to bind to the port but can't due to socket in use?
14:23 Ahlee though pretty sure minion sets the reuse flag
14:23 zerthimon Ahlee: restarting and killing processes doesn't help
14:23 Ahlee what versions?
14:23 sgargan is it possible to have salt log to two separate places i.e. /var/log/salt and also to syslog
14:24 zerthimon 2015.5.0+ds-1trusty1
14:24 clintberry joined #salt
14:24 Ahlee and minion has worked previously?
14:24 zerthimon and that /proc/.../tcp file is 4Gb... why is it reading it
14:26 Ahlee that's really, really odd.
14:26 Ahlee I'm out of ideas. My next step would be to bring the minion down, blow away /var/cache/salt/minion/* and restart
14:27 gmoro joined #salt
14:30 zerthimon Ahlee: did that ... no change
14:30 Ahlee debug logging enabled?
14:30 Ahlee if not, can it be?
14:30 zerthimon Ahlee: not enabled, I'll enable it
14:30 btorch hmm is there a problem having minions on 2014.7.1+ds-3lucid6 while the master is on latest ?
14:31 btorch there is no higher pkg available for lucid it seems
14:36 ange murrdoc: would you have an example ?
14:37 murrdoc sorry ange of  ?
14:37 murrdoc module.run ? http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html
14:37 ange module.run + cp.get_url
14:38 pas joined #salt
14:38 murrdoc https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html
14:38 ange that's slim docs
14:38 ange but thanks will found a way
14:38 ange *find
14:39 murrdoc its really easy , state_id: module.run: - name: cp.get_url - path: URL
14:39 murrdoc it will fail on non 200 status
14:40 ange well the fact that for instance I have to split name and url is odd compared to other use of similar calls (cmd.run for one)
14:42 irctc322 joined #salt
14:42 pas Hi. I upgraded salt-master from 2014.1.0 to 2015.5.0 (Lithium) which also upgraded PyYAML. Now when I try to yaml.load grains.items of a minion it fails with  yaml.scanner.ScannerError: mapping values are not allowed here. Has anyone else run into this?
14:44 iggy pas: I doubt many people yaml.load grains
14:46 epcim joined #salt
14:47 colttt joined #salt
14:48 ange still, would be nice for it to work if it's supposed to, no ?
14:49 iggy that wasn't the question
14:51 sgargan is it possible to configure the logging on a minion via its pillar information?
14:52 ange murrdoc: cool, worked, had to add the dest too but it does what I need; thanks
14:52 toanju joined #salt
14:53 sgargan can i have a logging.sls pillar where I change the log_file (to syslog) or the loglevel and have the minions pick this up when the pillars are refreshed?
14:53 murrdoc ange:  uh hmm
14:56 repl1cant joined #salt
14:56 joeto joined #salt
15:00 iggy sgargan: you can manage the minion config file via salt (see salt-formula f.ex.)
15:00 otter768 joined #salt
15:01 zerthimon Ahlee: from debug logs it looks to me a job status.master
15:04 desposo joined #salt
15:04 zerthimon Ahlee: so I ran salt-call status.master and that also took 100% cpu for 1 min or so
15:05 jalbretsen joined #salt
15:07 ange is there a good way to wait a couple of seconds for a process to spin up before proceeding with a cmd/state ?
15:08 ange http://docs.saltstack.com/en/latest/ref/states/all/salt.states.service.html I guess
15:09 sdm24 joined #salt
15:10 ange require: service.running: bind ?
15:10 dopesong joined #salt
15:11 repl1cant joined #salt
15:11 onorua joined #salt
15:12 iggy there's a difference between a service being "running" and actually being useful
15:15 ange ...
15:15 ange ok, let's be more precise
15:15 subsigna_ joined #salt
15:15 ange state A starts httpd daemon
15:16 ange state B needs to do a healthcheck on a url served by said httpd daemon
15:16 ange as you point out we can have a pid already but httpd daemon is not able to serve requests yet
15:17 iggy I don't think there's a generic way to handle that in salt
15:18 ange so, one could have a wait loop or a possible number of trials before considering it won't come to lif
15:20 aurynn this sounds like something your HA config should do, not salt?
15:21 ange could be
15:22 iggy there are probably quite a few ways to solve it
15:23 iggy I tend to lean more toward event driven everything when possible (server fires off a salt event when it's ready to serve, salt reactor does the rest, etc)
15:25 ange ah, interesting concept
15:26 ange iggy: do you any article / post detailing such a setup ?
15:26 ange too much coffee
15:26 ange iggy: do you have any article / post detailing such a setup ?
15:27 dfinn joined #salt
15:27 iggy uh... not really
15:27 rojem joined #salt
15:27 iggy I probably should though
15:27 iggy I guess I don't generally consider how much stuff I do that would probably be interesting to others
15:27 ange sounds like an interesting use of reactors and events
15:28 ange iggy: I started writing docs and articles about stuff I learn to first help my future self to remember, and at some point someone else will eventually points some possible improvements or use it
15:29 theologian joined #salt
15:29 ange and nowadays my team is quite thankful I write docs and wiki pages about what I build
15:29 ggoZ joined #salt
15:30 iggy too lazy to setup a website
15:30 ange a git repo
15:30 ange text files
15:36 denys joined #salt
15:39 zer0def joined #salt
15:42 c10 any ideas why "ssh_known_hosts.present" returns "Unable to receive remote host key"?
15:42 c10 host seems accessible if i ssh to the server and run "ssh -T"
15:42 Xiao joined #salt
15:43 fllr joined #salt
15:43 writtenoff joined #salt
15:49 ALLmightySPIFF joined #salt
15:49 mephx joined #salt
15:54 ecerulm joined #salt
15:54 nodens heya here
15:54 nodens anyone using syndic ?
15:55 nodens just configured a master and connected it on my master of master, added a minion on the local master pointing on itself
15:56 nodens accepted keys on local master and master of masters
15:56 nodens now, on master of masters, when I try a test.ping, i receive 3 responses
15:57 nodens with a grain.items, I get 3 responses as well, perfectly identical
15:57 nodens is it to be expected ?
15:58 ndrei joined #salt
16:01 numkem joined #salt
16:02 jamesog joined #salt
16:06 bandito joined #salt
16:06 edrocks joined #salt
16:07 Ahlee zerthimon: that's really, really wierd. Can you open a github issue for this?
16:07 zerthimon Ahlee: I'm writing it ATM, I know whats the problem
16:07 desposo joined #salt
16:07 Ahlee oh score
16:07 zerthimon Ahlee: it's a load balancer with a TON of tcp connections, and /proc/net/tcp file is HUGE
16:08 Ahlee ah
16:08 zerthimon Ahlee: status.master reads it, I'm not sure why
16:08 felixhummel joined #salt
16:08 bandito hi all, I'm having trouble duplicating the setup https://docs.saltstack.com/en/latest/topics/mine/index.html - trying to build a nagios config file that points at all minions. `salt '*' mine.get '*' network.ip_addrs` returns data, by the loop in the templated config file doesn't return anything
16:08 Ahlee i imagine it's trying to check if it's connected
16:08 Ahlee in a psuedo-naieve way
16:09 zerthimon Ahlee: on a 'small' server that file is a few K's
16:09 Ahlee right
16:09 zerthimon Ahlee: on a LB server I have it's 4GB
16:09 Ahlee not so small on a LB with hundreds of thousands of active tcp connections
16:09 thehaven_ joined #salt
16:09 bhosmer joined #salt
16:09 Ahlee wait, not hundreds, tens
16:09 Ahlee else you run out of ephemeral ports
16:09 Ahlee i don't know, networking hard
16:10 zerthimon Ahlee: yep
16:10 numkem joined #salt
16:10 giantlock joined #salt
16:11 sporkd2 Hey everyone, is it possible to have two states in a reactor sls?
16:12 iggy bandito: try .iteritems()
16:12 coval3nce joined #salt
16:12 Xiao joined #salt
16:13 bandito iggy: same result i'm afraid
16:13 nicmoski joined #salt
16:14 iggy bandito: just doesn't loop at all? no errors?
16:14 tkharju joined #salt
16:14 bandito iggy: that's correct - no errors, the rest of the file is created, however nothing from inside the loop appears in output
16:17 Bryson joined #salt
16:18 Bryson joined #salt
16:20 Bryson left #salt
16:21 Bryson joined #salt
16:23 rojem joined #salt
16:29 bandito iggy: i've created http://pastebin.com/vAuhyQn9 with a bit more information about the problem
16:29 iggy bandito: and the host you are testing on definitely shows output for salt minion mine.get '*' network.ip_addrs
16:30 bandito iggy: that's correct
16:30 iggy oh
16:30 iggy expr_form
16:30 iggy take that out
16:31 fllr joined #salt
16:31 bandito :D sweet!
16:31 bandito thanks iggy
16:32 scoates still having problems with `salt-call state.highstate` being way verbose. Not sure what I'm doing wrong. I want output like this: http://paste.roguecoders.com/p/0b4ac666896813c261eb46b4499d4e46.txt ; instead I get output like this: http://paste.roguecoders.com/p/17618a91686940e2f70560c3a4205644.txt … ideas?
16:33 clintber_ joined #salt
16:33 zerthimon Ahlee: could this problem be caused by having master_alive_interval parameter configured in minion's config ?
16:34 iggy scoates: I generally think of salt-call as a debugging tool
16:35 Ahlee zerthimon: I do not know enough about that code path to comment
16:35 ageorgop joined #salt
16:36 scoates iggy: calling `salt nodename state.highstate` from the master gives the same style of output
16:37 scoates it ends with http://paste.roguecoders.com/p/04b39a4cf670823ad0c2000e213bd786.txt ; I need to scroll way up to even see an error… and I don't seem to get a summary
16:37 KyleG joined #salt
16:37 KyleG joined #salt
16:37 iggy then you have something configured weird, my state.highstate output is just a list of states and their output... no debug/info/trace lines
16:37 zerthimon Ahlee: nope, I disabled it, and still the minion runs status.master every minute
16:37 amcorreia joined #salt
16:38 scoates iggy: so is mine when it goes right. Now it's going wrong. I suspect I *do* have something configured weird. Was hoping someone has seen/"fixed" this.
16:38 iggy scoates: right, use state_verbose: False to get rid of state output for states that didn't "change"
16:39 iggy but that looks like you have a different outputter set
16:39 iggy we used to use state_output: mixed
16:40 iggy it helped, but state_verbose: False obviates that
16:40 scoates thought I had that, but I guess it was on the minion. I get an actual error now, so that's helpful.
16:41 scoates I also do have `state_output: mixed`
16:41 iggy do you have output set to somehting other than nested?
16:41 iggy because your output looks like it's a different outputter
16:42 scoates I don't think so. Is there a way to ask salt to tell you its configuration? Perhaps something is different from what's in /etc/salt/master.
16:42 iggy config.get output
16:42 druonysus joined #salt
16:42 scoates returns empty
16:43 scoates http://paste.roguecoders.com/p/ac4e99d494b8183791eefb4cb2f0bc14.txt
16:43 rogst_ joined #salt
16:43 londo_ joined #salt
16:43 __alex_ joined #salt
16:43 cheus_ joined #salt
16:43 kivihtin joined #salt
16:43 iggy that'd be the minion side, not sure about the master
16:43 skepner_ joined #salt
16:43 tzero_ joined #salt
16:44 Antiarc_ joined #salt
16:44 edulix_ joined #salt
16:45 * scoates nods
16:45 MatthewsFace joined #salt
16:45 iggy grep -rin ^output /etc/salt
16:45 iggy that's the only way I can think
16:46 scoates ye; same. nothing. so should be defaults
16:46 scoates *yep
16:47 zerthimon Ahlee: I opened issue: https://github.com/saltstack/salt/issues/24758
16:47 iggy scoates: that's definitely not what my output looks like
16:48 Xiao joined #salt
16:48 scoates yeah; mine doesn't always, either. It's like something is failing and causes the output to change.
16:49 Ahlee zerthimon: fantastic issue, thank you for providing the details. You might want to append the output of salt --versions
16:50 zerthimon Ahlee: will do
16:53 tpaul joined #salt
16:54 scoates now (with state_verbose: false) all I get on highstate (from the master) is http://paste.roguecoders.com/p/2b4c25b4154e5b154668e9721155f000.txt … if I'm not mistaken, there's no indication of what's actually failing there. /-:
16:55 theologian joined #salt
16:56 tpaul Using salt-ssh for the first time, attempted to ping and recieved "Host key verification failed.". Does the server need to be added to my known_hosts file before using salt-ssh or is there a way to handle this through salt? Thanks!
16:56 scoates I'm using salt 2015.5 from Debian Jessie… which is probably the *actual* problem… but bootstrap + systemd is broken on Jessie. Fun times.
16:57 rojem joined #salt
16:57 iggy scoates: start disabling states until you find the issue
16:58 scoates yeah. that's the strategy. much pain. not sure how to help make that better, either.
16:58 scoates thanks, iggy
16:58 iggy when you find the problem, you might file a bug
16:59 iggy we don't like tracebacks (although sometimes they are unavoidable)
16:59 bhosmer joined #salt
17:00 jamesog joined #salt
17:00 scoates I'll try, yes. It's often difficult to distill this kind of problem down into a test case without spending many hours on it (in addition to the hours wasted finding the problem in the first place)
17:01 otter768 joined #salt
17:05 skepner_ joined #salt
17:06 ALLmightySPIFF joined #salt
17:07 mou joined #salt
17:08 bradthurber joined #salt
17:10 drawsmcgraw joined #salt
17:11 bradthurber Hi. Salt milestones name references in the salt docs are driving me crazy (feature blah will be/was deprecated in Salt Carbon, etc). I found this milestone page https://github.com/saltstack/salt/milestones, but is there a cross-reference to released versions?
17:14 iggy I printed out a periodic table and just wrote the version numbers on it
17:15 druonysus joined #salt
17:15 druonysus joined #salt
17:19 forrest joined #salt
17:20 Edgan joined #salt
17:21 scoates iggy: FWIW, I've isolated it to a big #!py state; I'm sure I made a mistake in there. Going to be difficult to track it down to something reportable, though, I bet.
17:22 conan_the_destro joined #salt
17:26 numkem joined #salt
17:27 zircote joined #salt
17:28 numkem joined #salt
17:31 supersheep joined #salt
17:31 bhosmer joined #salt
17:32 c10 joined #salt
17:33 tomh- joined #salt
17:35 c10 joined #salt
17:36 julez joined #salt
17:36 ingslovak joined #salt
17:39 hal58th joined #salt
17:39 hal58th_ joined #salt
17:39 skepner_ joined #salt
17:41 bradthurber joined #salt
17:41 sdm24 does anyone familiar with apache know if there is any difference if I have salt file.manage a host.conf file in both the sites-enabled and sites-available directory, compared to managing just sites-available and have salt create the symlink in sites-enabled?
17:41 sdm24 Will they both work the same?
17:43 iggy it's more standard to see links in sites-enabled
17:43 nkuttler sdm24: they should
17:44 iggy don't know whether that's a + or - for using them though
17:44 bradthurber trying to get salt-cloud configured with rackspace as the cloud provider. I am getting error "<LibcloudError in None 'OpenStack instance must have auth_url set'>". I see identity_url in the examples, but not auth_url. Even if I add auth_url to the .conf file I still get the error. I'm using latest apache-libcloud. Is auth_url possibly a new paramet
17:44 bradthurber er required?
17:44 nkuttler well. the whole sites-enabled/available is not upstream anyway, is it?
17:44 scoates debianism
17:44 nkuttler yeah
17:45 iggy pretty sure more than debian uses that at this point
17:45 iggy but either way... ewww, apache
17:45 Heartsbane joined #salt
17:45 Heartsbane joined #salt
17:45 nkuttler all debian derived distros? what about centos?
17:45 sdm24 thanks iggy and nkuttler. For me, its easier to just write two file.managed IDs than to worry about creating symlinks
17:45 nkuttler sdm24: fwiw, i do the same thing
17:45 murrdoc terribad
17:46 murrdoc the recommendation is make files in sites-available and symlink into sites-enabled
17:46 murrdoc from nginx itself , i think
17:46 nkuttler we're not talking nginx though
17:46 nkuttler wait.. what
17:46 murrdoc :D
17:47 sdm24 nkuttler: and it works for you? perfect. I guess  I need to actually create an index.html to test it out
17:47 nkuttler yeah. anyway. i see no point in using symlinks. if you use config management, managing config files is what you do. you don't need to add another layer (symlinks) to manage config files
17:47 hunmaat if you dont want to make a symlink, then dont create the file in sites-available
17:47 hunmaat its really misleading to have a copy there
17:48 hunmaat and it's not used for anything else than a2en/dissite
17:48 nkuttler yeah, don't create two copies
17:48 sdm24 Ok thanks
17:48 hunmaat but having a dead copy there seems terrible
17:49 sdm24 I wasn't sure if it needed to be there. I don't know much about apache : /
17:49 murrdoc terribad
17:50 murrdoc (not terribad)
17:50 murrdoc but yes bad
17:50 murrdoc you dont one changed and the other not changed
17:50 hunmaat the only point of the directory pair is to make sites (modules, confs) easily switchable
17:50 murrdoc hunmaat:  speaks truth
17:50 ndrei_ joined #salt
17:50 linjan joined #salt
17:56 _mel_ joined #salt
17:58 matthew-parlette joined #salt
17:59 darknight87 joined #salt
18:00 toanju joined #salt
18:01 scoates iggy: I managed to distill it down to something pretty small (top.sls + two .sls files); I now see an error in the log, but there's a traceback. Should all traces be reported? Or is that expected?
18:01 iggy depends
18:01 iggy they can't catch _every_ issue in the universe, but if it's something someone is likely to hit, go for it
18:01 scoates http://paste.roguecoders.com/p/29410b78159b75eea4b070555201796c.txt
18:02 scoates recursive requisite… would be nice to see it not dump, though, I think
18:03 iggy yeah, I'd file (search first, there might be something already)
18:03 * scoates nods
18:03 scoates thanks
18:03 epcim joined #salt
18:03 cruatta joined #salt
18:03 iggy I've hit that before (using variables as requisites it can be easy to miss...)
18:04 scoates I don't see an issue. Will report. Thanks for the push and the help.
18:04 dopesong joined #salt
18:06 forrest_ joined #salt
18:07 ALLmightySPIFF joined #salt
18:14 scoates https://github.com/saltstack/salt/issues/24761
18:15 sieve joined #salt
18:15 scoates now… what was I doing? (-:
18:16 subsignal joined #salt
18:16 DammitJim joined #salt
18:17 skepner_ joined #salt
18:19 swa_work joined #salt
18:20 ajw0100 joined #salt
18:25 soren joined #salt
18:29 bradthurber joined #salt
18:30 denys joined #salt
18:32 linjan joined #salt
18:34 baweaver joined #salt
18:34 darknight87 joined #salt
18:35 dopesong_ joined #salt
18:38 aron_kexp joined #salt
18:40 mens joined #salt
18:41 ajw0100 joined #salt
18:43 johann joined #salt
18:44 tpaul sdm24: I can tell you how we use apache+salt on rhel which doesn't have the sites-enabled stuff, but I don't know if that will help you
18:44 sdm24 tpaul: might as well. Is it wildly different?
18:45 aboe joined #salt
18:45 johann Hey guys - I'm running salt 2014.1.5 with the default zero-mq. This are in a weird state where sometimes jobs succeed on a minion, but often don't. Is there any way to reset/or flush zero-mq?
18:47 Deevolution Anyone have any experience with creating custom renderers?  I'm having trouble getting Salt to load them and/or it throws errors when running them.  Even if I create a renderer that returns a static set of data.
18:47 NV joined #salt
18:48 tpaul sdm24: We have a customized version of the default global apache config managed by salt (/etc/httpd/conf/httpd.conf) with an Include conf.d/*.conf directive at the end, and then all web applications using apache install a vhost config into /etc/httpd/conf.d/
18:48 stoogenmeyer joined #salt
18:48 johann Also - ZMQ: 4.0.4
18:49 sdm24 oh I see. so apache just needs to load vhost.conf, it doesn't have to be in sites-enabled? That's just where apache looks to by default?
18:51 tpaul I'm not 100% sure about it being an apache default, but on the rhel version placing a virtual host config into httpd/conf.d/ "enables" the site (after a config reload of course)
18:51 sdm24 cool, thanks
18:51 tpaul np
18:54 tpaul Anyone here using salt-ssh? I'm trying it out for the first time and it doesn't appear to be sending the password specified in the roster
18:55 tpaul I turned on trace logging and there is a line that says "executing command: " and the password is not in the command, if I run the command as listed in the log it works as long as I type in the pass when prompted
18:55 mikepea joined #salt
18:56 matthew-parlette joined #salt
18:59 bhosmer joined #salt
19:02 otter768 joined #salt
19:02 peters-tx joined #salt
19:03 bradthurber joined #salt
19:07 ALLmightySPIFF joined #salt
19:16 spookah joined #salt
19:16 davisj hrmmm... How do I get to pillar data for a minion via python? I know I can do salt.loader.grains(__opts__) for grains, but don't see the pillar equivolent.
19:17 murrdoc gist ?
19:17 vagnerd joined #salt
19:17 davisj murrdoc: not much to paste, just looking at http://docs.saltstack.com/en/latest/ref/clients/index.html
19:18 bluenemo joined #salt
19:18 bluenemo joined #salt
19:21 hybridpollo joined #salt
19:22 swa_work joined #salt
19:22 ageorgop joined #salt
19:24 julez joined #salt
19:25 pcn Can I render pillars if I want to have jinja templates in them that query grains?
19:25 Edgan joined #salt
19:26 sdm24 https://gist.github.com/sdm24/ad5a9517b4818b6c326f like that pcn?
19:27 pcn sdm24: yeah, I'm having issues with that on my end.
19:28 sdm24 works for me on the minions targetted in /srv/pillar/top.sls
19:32 dopesong joined #salt
19:32 nicmoski joined #salt
19:33 nicmoski Is it possible to import from existing custom modules (located in _modules) without doing this: sys.path.append('/var/cache/salt/minion/extmods/modules')?
19:34 dopesong joined #salt
19:35 Tod-Work joined #salt
19:39 ajw0100 joined #salt
19:40 thedodd joined #salt
19:43 davisj dir(__salt__)
19:43 davisj bah
19:43 pcn sdm24: it seems that with a jinja template, I can do {% else %} but {% else  -%} is a rendering error.
19:46 pcn with salt-ssh, where do errors get logged so I can troubleshoot?  All I see is:
19:46 pcn _errors:
19:46 pcn - Rendering SLS 'cassandra' failed. Please see master log for details.
19:46 pcn That's not very helpful when I don't have a master
19:48 xGoG13 joined #salt
19:48 xGoG13 hi
19:49 davisj Ah, to answer my own question: caller = salt.client.Caller() \ caller.function('pillar.items')
19:50 xGoG13 is possible inject jinja data into cloud porfiles?
19:50 baweaver joined #salt
19:50 xGoG13 *profiles
19:51 baweaver joined #salt
19:52 rdutch joined #salt
19:53 murrdoc joined #salt
19:54 sdm24 pcn, it might be because -%} will remove whitespace after the jinja, which might break your YAML format
19:54 pcn That is interesting, and makes some sense.
19:55 xGoG13 i wanted inject formula to add unique id on create vm cloud, example:  {{ salt['cmd.run']('date +%s') }}
19:56 pcn Related question (to me): can I call custom grains from inside a pillar?
19:56 pcn E.g. if I use the ec2_info grain from the contrib repo, can I access its data from within a pillar?
19:56 roder joined #salt
19:57 rdutch pcn: yes you can
19:57 Zachary_DuBois joined #salt
19:58 cruatta_ joined #salt
19:59 soren joined #salt
19:59 armguy joined #salt
20:00 pcn rdutch: how would you debug an issue relating to that?  I'm using salt-ssh to try this out, and I can't figure out where errors are logged
20:00 cruatta_ joined #salt
20:01 roder Can someone verify for me that salt.states.bower is in the 2015.5.2 docs, but not in the 2015.5.2 release? Also 2015.5.0 is what is installed with bootstrap.saltstack.com (at least for ubuntu stable) currently?  Is that all what should be expected?
20:02 dopesong_ joined #salt
20:04 c10 joined #salt
20:05 rdutch pcn: i would look on the master logs or start master in foreground wkith -l debug
20:05 babilen roder: It is in neither: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.bower.html (vs. http://docs.saltstack.com/en/develop/ref/states/all/salt.states.bower.html which is the development branch)
20:06 babilen And yes, 2015.2.0 is currently the latest packaged version for Ubuntu
20:07 roder ok cool.  thanks - I also just realized that I misread which docs I was on (I was on develop, not 2015.5.2) thanks for pointing that out :)
20:07 roder i thought i was going crazy
20:07 skepner_ joined #salt
20:07 babilen Just because you have been reading the develop docs doesn't mean that you are not going crazy
20:08 babilen >:-)
20:09 ALLmightySPIFF joined #salt
20:09 roder it's true.
20:11 pcn rdutch salt-ssh at the moment, so I don't think there are master logs
20:11 pcn or a master in the background or foreground
20:12 tiadobatima joined #salt
20:12 pcn Interesting: if I have a custom grain that creates a dictionary with an integer key, accessing that key from jinja/salt requires a string, not an int
20:13 toddnni joined #salt
20:13 s_kunk joined #salt
20:16 ecerulm joined #salt
20:18 ggoZ joined #salt
20:18 fannet joined #salt
20:19 fannet hey everyone - I have a question on a multi-master setup. If I issue a high state from the master on a bunch of minions - will all those minions use the master that originated that job or will they use any of the available masters?
20:19 s_kunk joined #salt
20:21 fannet hey everyone - I have a question on a multi-master setup. If I issue a high state from the master on a bunch of minions - will all those minions use the master that originated that job or will they use any of the available masters?
20:22 armguy joined #salt
20:22 baweaver joined #salt
20:23 toddnni_ joined #salt
20:24 forrest joined #salt
20:25 bhosmer_ joined #salt
20:26 andrew_v joined #salt
20:28 Slimmons joined #salt
20:29 sporkd2 can you not use set = mine.get in an sls ?
20:30 sporkd2 I have a reactor state that I'd like to get the mine grains off of a minion and pass that to the orchestrate sls but it doesn't seen to want to execute
20:31 dalexander joined #salt
20:31 bhosmer_ joined #salt
20:31 dyasny joined #salt
20:31 sdm24 sporkd2 you need to use a for loop instead of set
20:32 sdm24 https://docs.saltstack.com/en/develop/topics/mine/index.html#example
20:32 dyasny joined #salt
20:32 c10 any ideas why I would be getting "Permission denied for host sever1, do you want to deploy the salt-ssh key? (password required): " with salt-ssh? (tested ssh to the roster server, and it works)
20:32 sporkd2 thanks sdm24 I'll try that
20:33 baweaver joined #salt
20:36 murrdoc joined #salt
20:38 jalbretsen joined #salt
20:38 mou joined #salt
20:40 ALLmightySPIFF joined #salt
20:41 DammitJim joined #salt
20:44 hal58th_ joined #salt
20:44 dopesong joined #salt
20:45 KyleG joined #salt
20:45 KyleG joined #salt
20:45 premera joined #salt
20:47 Antiarc joined #salt
20:50 murrdoc joined #salt
20:52 cruatta joined #salt
20:54 cruatta_ joined #salt
20:54 fannet should a multi-master setup use a shared job cache?
20:56 c10 so nobody has any idea about the salt-ssh password required issue?
20:57 rdutch left #salt
20:57 fannet sounds like it didn't accept your key
20:58 whytewolf c10 -l debug just after salt-ssh
20:59 cberndt joined #salt
20:59 ctolsen joined #salt
21:01 armguy joined #salt
21:01 c10 whytewolf: http://pastebin.com/DUZ2SHrM
21:02 c10 (ip was replaced)
21:02 giantlock joined #salt
21:03 otter768 joined #salt
21:03 renoirb joined #salt
21:03 sgargan joined #salt
21:08 tpaul c10: I was running into the exact same thing today, I had manualyl work around, what platform/version are you on?
21:09 bhosmer joined #salt
21:09 c10 osx, latest version
21:09 c10 10.10.3
21:10 cruatta joined #salt
21:10 c10 the weird thing is i'm getting this only on one of 3 boxes
21:10 c10 with the same os
21:11 c10 tpaul: how did you work around it
21:12 tpaul I used ssh-copy-id to copy the salt-master generated public key to the server
21:13 tpaul For me it's OS specific it only seems to be happening on NetBSD 7, Fedora seems to work ok
21:13 jm_ joined #salt
21:14 jm_ Does salt-minion need to run on the salt-master host machine?
21:15 tpaul jm_: No, but it can
21:16 jm_ What would be the typical reason to do it?
21:19 c10 damn, this bugs me so much
21:19 c10 :)
21:19 tpaul One of the reasons I use salt is to enforce some security policies across all servers (firewall rules, selinux settings, etc), the master is not immune to these settings
21:20 tpaul c10: did that work?
21:20 jonher937 jm_: To manage the master (user accounts, firewall, but also config for the master itself with version history in git)
21:21 jm_ Thank you, that makes sense
21:22 ahammond is there a way to test my targeting that doesn't contact minions? I'm using E and L type targeting, no grains or anything that would actually require a minion to process
21:22 c10 tpaul: nope, "All keys were skipped because they already exist on the remote system."
21:23 tpaul That's frustrating
21:24 c10 it's annoying because i know not os related, but i can't even think of ways to debug it
21:24 tpaul have you tried running salt-ssh -l all ... and sifting through the output?
21:25 tpaul I had a fun time with that today :)
21:34 Momo_ joined #salt
21:38 baweaver joined #salt
21:38 c10 tpaul: figured it out. i needed to add the key to ssh-agent
21:40 righdforsa joined #salt
21:41 toddnni joined #salt
21:42 MatthewsFace joined #salt
21:44 ioc joined #salt
21:44 ioc hi all
21:46 sgargan joined #salt
21:46 ioc can you help me
21:46 ioc from hash?
21:46 murrdoc joined #salt
21:47 ioc please
21:50 supersheep joined #salt
21:55 cruatta joined #salt
21:56 Nazca__ joined #salt
21:57 rap424 joined #salt
21:58 badon joined #salt
21:59 kula joined #salt
21:59 rojem joined #salt
22:00 dlott joined #salt
22:04 edrocks joined #salt
22:06 righdforsa hi. is there a known issue with "require" for user and group states? Getting an error that '__prereq__' is an invalid keyword argument for 'user.present'
22:08 badon joined #salt
22:10 vagnerd joined #salt
22:17 sgargan joined #salt
22:17 ranomore joined #salt
22:20 epcim joined #salt
22:22 hemphill joined #salt
22:24 sgargan joined #salt
22:25 murrdoc has anyone done anything to perf their salt master
22:26 murrdoc i have the limits set to 4096 and the sysctls from the faw page
22:26 murrdoc want to know if there is other stuff
22:26 murrdoc that anyone can recommend
22:27 murrdoc like i dint know http://docs.saltstack.com/en/latest/topics/tutorials/intro_scale.html existed
22:30 ahammond I'm running a state across a lot of machines and am really only interested in problems. Is there an outputter or returner I should be using to make this easier to spot?
22:31 bitwise__ joined #salt
22:32 jm_ I sent a sigkill to an unresponsive salt-master running in the foreground. It seems to periodically hang here: File "/usr/lib/python2.7/dist-packages/salt/fileserver/__init__.py", line 48, in check_file_list_cache
22:32 jm_ cache_stat = os.stat(list_cache)
22:33 jm_ Should I clear the cache?
22:36 spookah joined #salt
22:36 spookah joined #salt
22:37 sgargan joined #salt
22:41 righdforsa ah, nevermind my question...looks like it was a "prereq" problem, and seems to be resolved with a tweak to state
22:41 Slimmons if i want to apply a part of a state (and only a single part of a state, like  just a git.latest) to a minion that isn't in the group of minions for that state (because I don't want that state to apply to that minion except for one small part), could I do anything other than circling every part of the state in conditionals like {% if grains['id'} == 'singleMinion' %}?
22:42 Slimmons I'm basically wanting a way to say this git.latest, is for this one minion, and this is the only thing that should apply to this minion
22:42 Slimmons and the rest of it goes to all the other minions
22:42 ALLmightySPIFF joined #salt
22:43 smcquay joined #salt
22:44 smcquay joined #salt
22:44 pimpind joined #salt
22:44 sgargan joined #salt
22:45 smcquay joined #salt
22:49 ctolsen joined #salt
22:50 Slimmons I think what I'm wanting can be accomplished by just adding a separate entry for that minion in the top.sls file, and JUST doing what I want there, and then running the next group of minions.  I just wasn't sure if something like this existed, or if I was going about it the wrong way.
22:50 Sage|4U joined #salt
22:51 TheScriptSage Hey folks.  Any chance of getting this today? https://github.com/saltstack/pepper/issues/33
22:51 TheScriptSage I know we can install directly from github with pip, but I need it for something that pulls from pypi.
22:54 MatthewsFace joined #salt
22:54 iggy TheScriptSage: I'm not sure any of the pepper devs hang out in here
22:54 TheScriptSage Damn.
22:55 iggy whiteinge is in here sometimes, but not today
22:55 TheScriptSage Ah, that's who I was looking for.
22:56 dlott joined #salt
23:03 bhosmer_ joined #salt
23:04 otter768 joined #salt
23:07 Slimmons iggy, or anybody else got a sec to help me with what I think is a simple problem?
23:07 iggy just ask
23:07 Slimmons it's confusing as hell when I try to type it all out at once :)
23:07 Slimmons alright, so I have a large group of servers, and I want a repo added to them.
23:07 Slimmons so
23:07 Slimmons git repo*
23:08 Slimmons I've created a minion on my master, so that I can clone the repo to the master, and then have the files served out to the minions
23:08 Slimmons my issue is, that the state is pretty big, and advice I was given earlier was to use conditionals to say what parts of the state applied to the minion on my master
23:09 Slimmons so something like {% if grains['id'] == 'mymastersminion' %}
23:09 Slimmons then put the git.latest in that
23:09 Slimmons the problem that iv'e hit (that i'm certain there's a simple solution for
23:09 iggy that was one possible solution that was mentioned
23:09 Slimmons ) is that i'd have to go through this huge state and put conditionals around everything
23:09 Slimmons yeah, and i'm sure that's not the only one
23:10 Slimmons I just thought that one seemed pretty good (originally)
23:10 Slimmons and there's nothing wrong with it, i'm sure i'm just implementing it incorrectly, or I'm not sure about a feature that salt has (which has been the case during every other problem i've had)
23:10 Slimmons so since I don't want to write a conditional for every single init.sls file I have, specifying that it applies to everything but my master's minion
23:11 Slimmons I was hoping there was another solution, or a way to logically say  the git.latest only applies to that one minion, and every other section of the state should ignore that minion
23:12 iggy gist your top file and maybe some of the states
23:14 Slimmons groupname:
23:15 Slimmons 'groupnameTest*':
23:15 Slimmons - postgresql
23:15 iggy gist
23:15 iggy not paste it in here
23:15 Slimmons <----noob
23:15 Slimmons gist it into another site, and link it
23:15 Slimmons or way to gist into irc idk about?
23:15 iggy Please use https://gist.github.com for code, don't paste directly into the channel
23:15 iggy from the /topic
23:16 iggy and gist supports multiple files (and revisiions)
23:16 Slimmons awesome, thanks, i'll do that
23:18 bluenemo joined #salt
23:18 monkey66 joined #salt
23:20 bfoxwell joined #salt
23:21 sambev joined #salt
23:22 baweaver joined #salt
23:23 mosen joined #salt
23:24 Slimmons https://gist.github.com/johnsimmons/2945a0eb5eeaf34d5f2c
23:24 Slimmons I would look at the environment.sls first, an dread the comments
23:25 Slimmons i appreciate the help.
23:25 iggy does that even work?
23:25 iggy your tipe file looks all kinds of wrong
23:26 iggy *top file
23:26 Slimmons yeah, it works (except I haven't tried it with the actual mastersminion added'
23:27 Slimmons adding mastersminion into the top file, and the conditionals in environment.sls are the only things that i haven't tested
23:27 iggy in your top file, groupname isn't a group, it's a saltenv
23:27 Slimmons yeah, i just quickly named it groupname.  It's not actually a group
23:28 iggy L@ implies that you are doing compound matching, but you have no `- match: compound` on the next line
23:28 Slimmons oh, i thought the L@ was a list of minions
23:28 Slimmons well, I just added that quickly, so show you what I was thinking, I didn't test that part.  I guess I didn't understand how the L@ worked correctly
23:29 Slimmons i had just done it before with node groups
23:29 Slimmons fyi, i'm not using nodegroups on this master
23:29 iggy so... you don't want postgres or any of the other stuff in your top file to install on your master
23:29 iggy correct?
23:30 druonysus joined #salt
23:30 druonysus joined #salt
23:30 conan_the_destro joined #salt
23:30 Slimmons correct, except for the one part in environment.sls
23:30 iggy so take the master out of that
23:31 iggy take the master specific part of the environment.sls and put it in a separate file (masterenvironment.sls or whatever)
23:32 iggy below your main group, add: 'master':\n  - masterenvironment
23:32 iggy well
23:32 Slimmons then i'd have to run two state.highstates right?
23:32 iggy so you actually need to put that above that group
23:32 Slimmons yeah
23:32 iggy then salt will do the master part first, then the other bits after
23:33 Slimmons wouldn't i then have to run "salt 'master' state.highstate"
23:33 Slimmons then salt 'minions' state.highstate?
23:33 Slimmons I apologize, this is something I must have missed while reading through the documentation, I'm sure this is super basic stuff
23:34 iggy nah, salt operates top down by default
23:34 iggy so if you order your top file correctly, everything should take place in the proper order
23:35 iggy I just commented on your gist
23:36 Slimmons ok, i changed the gist to represent what you meant, and I think i got it.
23:36 Slimmons so, my break in logic was.....I thought if you ran "salt 'groupnameTest1' state.highstate"    That it would only run the environment specified below the groupnameTest*
23:37 Slimmons and salt 'mastersminion' state.highstate, would then ONLY run what's under mastersminion
23:37 mou joined #salt
23:38 kevin-wk 'groupnameTest1' and 'mastersminion', in the context of those commands, are the targets
23:38 kevin-wk the top.sls file tells salt what to do for those targets
23:39 sgargan joined #salt
23:41 Slimmons right, but if both of them are targeted (i'm doing stuff with both), i didn't think it would work
23:41 kevin-wk it'll do the right thing for all of them
23:41 kevin-wk you can just "salt '*' state.highstate" if you want to highstate everything
23:42 kevin-wk each one would evaluate the top.sls in the base environment to figure out which state files to apply
23:43 kevin-wk in the case of the top.sls iggy added to your gist, the minion that matches 'mastersminion' would apply masterenvironment and minions that match 'L@groupnameTest*' would do all those other states
23:44 kevin-wk and if the minion didn't match either of those things nothing would happen
23:44 Slimmons that's awesome, and exactly what I needed.  Thanks to both of you
23:44 Slimmons thanks for spending time on noobs :)
23:44 kevin-wk np
23:48 NV joined #salt
23:49 c10 joined #salt
23:49 amcorreia joined #salt
23:51 sporkd2 hey guys, this is hair pulling for me. I am using a reactor state to call the runner.state.orchstrate and executing an orchestration state. I need to grab some mine data along the way but no matter where I put my mine.get for loop is does not render. If I put the same for loop in a managed file the orchestration sls renders is perfectly. Am I missing something?
23:53 desposo joined #salt
23:53 iggy Slimmons: just make sure everything is in your top file in the order it needs to occur and you should be fine with salt '*' state.highstate
23:53 otter768 joined #salt
23:54 Slimmons thanks iggy, will do
23:54 iggy sporkd2: I'd search/file a bug... mine functions should be available in the master context (imo)
23:55 iggy sporkd2: as a workaround, the mine should work in states, so just have your orch call states instead of functions directly
23:56 moos3 joined #salt

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