Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-07-06

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

All times shown according to UTC.

Time Nick Message
00:03 quasiben joined #salt
00:14 LV-426 besides jinja based approaches like {% if grains['fqdn'] == 'foo' %}, how can I perform minion targeting from within sls modules?
00:18 lala joined #salt
00:25 Bryson joined #salt
00:27 phx__ joined #salt
00:27 Arendtse1 joined #salt
00:27 zmodemz LV-426: what about targeting in the top.sls?
00:28 sunkist1 joined #salt
00:28 lynxman joined #salt
00:29 _gothix_ joined #salt
00:29 crd joined #salt
00:30 rome_390 joined #salt
00:30 finnzi joined #salt
00:31 georgemarshall joined #salt
00:32 aqua^c joined #salt
00:34 GothAck joined #salt
00:38 borgstrom joined #salt
00:41 otter768 joined #salt
01:00 mrbigglesworth joined #salt
01:00 heewa left #salt
01:02 bfoxwell joined #salt
01:13 mapu joined #salt
01:13 toastedpenguin joined #salt
01:14 toastedpenguin joined #salt
01:16 mapu left #salt
01:24 beauby joined #salt
01:36 mrbigglesworth joined #salt
01:37 madpenguin joined #salt
01:42 beauby joined #salt
01:45 aqua^c joined #salt
01:50 skullone LV-426 is an ugly planet, a bug planet
01:58 markm_ joined #salt
02:06 beauby joined #salt
02:07 kevin-wk left #salt
02:11 malinoff joined #salt
02:13 bastiandg joined #salt
02:26 mrbigglesworth joined #salt
02:28 beauby joined #salt
02:36 stanchan joined #salt
02:38 aqua^c joined #salt
02:38 aqua^c joined #salt
02:40 yomilk joined #salt
02:41 napsterX joined #salt
02:41 hasues joined #salt
02:41 hasues left #salt
02:55 kal0pr joined #salt
02:58 beauby joined #salt
03:01 kphannan joined #salt
03:01 zz_cro joined #salt
03:10 beauby joined #salt
03:17 otter768 joined #salt
03:19 Bryson joined #salt
03:26 favadi joined #salt
03:27 timoguin joined #salt
03:33 jaybocc2 joined #salt
03:35 lictor36 joined #salt
03:45 Guest89_ joined #salt
03:50 clintberry joined #salt
03:53 clintber_ joined #salt
03:55 jaybocc2 joined #salt
03:57 kal0pr joined #salt
04:12 stanchan joined #salt
04:17 desposo joined #salt
04:20 scoates joined #salt
04:26 jaybocc2 joined #salt
04:30 sirtaj left #salt
04:35 bones050 hi guys, I'm using Gitfs in my salt environment. But I'm in a situation where I need salt to gather some pillar information from a new host and add that to a configuration file that is managed via salt. If I edit this directly on the host without updating gitfs, then the file will be overwritten later - as the gitfs holds the 'gold' copy. Is there any way to update this?
04:40 jaybocc2 joined #salt
04:44 ramaseshan joined #salt
04:49 ramteid joined #salt
04:52 rdas joined #salt
04:52 napsterX joined #salt
05:00 markm__ joined #salt
05:03 catpiggest joined #salt
05:04 mrbigglesworth joined #salt
05:17 otter768 joined #salt
05:20 jaybocc2 joined #salt
05:21 tkharju joined #salt
05:28 cilkay Hello. I have masterless minions to which the people who will be running those VMs will have no shell access at all. I want to be able to provide them with feedback on the console as the various scripts are being run as the machine auto-configures or auto-updates itself. I know I can output messages to /dev/tty. How can I output simple, concise messages to the users? I don't want to output the entire output that I see if I salt-call highstat
05:28 cilkay e.
05:28 cilkay That's way too much detail for the users.
05:29 napsterX_ joined #salt
05:33 stephanbuys joined #salt
05:44 LotR joined #salt
05:46 Mate cilkay: try calling with --state-output=mixed
05:47 cilkay I've never seen that one. I'll try it. Thanks!
05:47 Mate it will give only a success/clean/fail info about each state
05:48 Mate and details if failed
05:49 cilkay It gives me everything + the summary.
05:49 gcfhvjbkn joined #salt
05:50 cilkay The summary is shown at the end. I'd like to be able to set my own messages for each of the various scripts. I saw something about that in the docs but can't remember where.
05:51 Mate scripts?
05:51 cilkay I call them scripts for lack of a better word. The .sls files.
05:51 xmj left #salt
05:52 TyrfingMjolnir joined #salt
05:53 napsterX joined #salt
05:53 Mate the state failes have no real role after the states to be applied are collected
05:53 Mate so they arent summed per state file
05:54 cilkay That explains why things aren't necessarily applied in the order in which they appear in top.sls
05:59 Mate you can be only sure if requirements (or static order values) are set
06:02 chiui_ joined #salt
06:03 evle joined #salt
06:05 seafloss joined #salt
06:07 colttt joined #salt
06:09 mosen joined #salt
06:12 dopesong joined #salt
06:13 marnom joined #salt
06:15 kal0pr joined #salt
06:21 hieund12 joined #salt
06:27 marnom I changed agent connection method (over internet instead of VPN, more bandwidth that way) and now salt-run manage.present is no longer seeing the minions in that subnet as present, though I can test.ping/highstate them fine :|
06:29 flyboy joined #salt
06:30 marnom I already tried redeploying the VMs and re-accepting keys on the master, restarting master etc.. no go. Also running latest 2015.5.2 release on debian. Any suggestions maybe?
06:31 stoogenmeyer joined #salt
06:35 KermitTheFragger joined #salt
06:43 stephanbuys joined #salt
06:44 stephanbuys1 joined #salt
06:50 kal0pr joined #salt
06:55 illern joined #salt
06:55 thalleralexander joined #salt
06:57 markm__ joined #salt
07:05 madpenguin joined #salt
07:07 AndreasLutro joined #salt
07:12 crocket joined #salt
07:12 crocket I get an error, "Rendering SLS ssh failed, render error: '__env__'" when I execute state.highstate.
07:12 crocket The error emerges from a pydsl state.
07:13 crocket What does it mean?
07:13 phix joined #salt
07:13 ndrei joined #salt
07:14 jaybocc2 joined #salt
07:14 gcfhvjbkn joined #salt
07:18 otter768 joined #salt
07:20 kawa2014 joined #salt
07:26 lionel joined #salt
07:30 jeddi joined #salt
07:33 AndreasLutro crocket: can you get the full stack trace/
07:33 AndreasLutro ?
07:34 crocket http://dpaste.com/008N7KA
07:34 crocket AndreasLutro, ^^
07:35 ndrei joined #salt
07:36 AndreasLutro is that all of it?
07:36 crocket Yes
07:37 crocket oh
07:37 crocket AndreasLutro, There's more
07:37 crocket http://dpaste.com/2QGBKW6
07:37 crocket AndreasLutro, ^^
07:37 AndreasLutro cheers
07:38 crocket The same code leads to a traceback on my server wihle it works fine on my desktop.
07:38 linjan joined #salt
07:38 crocket It is a heisenbug.
07:38 AndreasLutro what salt version?
07:38 crocket salt-call 2015.5.2 (Lithium)
07:39 crocket on my desktop and my server
07:40 AndreasLutro crocket: you should report an issue on github. for now I can see that the error is caused by having two states with duplicate IDs so maybe you can look for that
07:41 crocket AndreasLutro, How did you infer that two states had the same ID from the traceback?
07:41 stephanbuys joined #salt
07:41 khebbie joined #salt
07:41 AndreasLutro crocket: looked at the stacktrace, found the line where the error is occuring https://github.com/saltstack/salt/blob/v2015.5.2/salt/state.py#L2854
07:41 AndreasLutro as you can see, it's erroring as it's trying to append an error to the highstate run
07:42 stephanbuys left #salt
07:42 crocket http://dpaste.com/03HYQSH
07:43 crocket AndreasLutro, ^^ is the problematic code.
07:43 ljt joined #salt
07:45 AndreasLutro crocket: maybe you have other state files with a state the same ID as a user in your system? hard to guess
07:46 crocket Ah
07:46 crocket AndreasLutro, I defined '/root/.ssh' in two separate files.
07:47 crocket The traceback was misleading...
07:49 jaybocc2 joined #salt
07:51 fivmo joined #salt
07:55 phix joined #salt
07:57 sectionme joined #salt
08:00 al joined #salt
08:01 impi joined #salt
08:02 keimlink joined #salt
08:02 phix joined #salt
08:05 chiui_ joined #salt
08:08 crocket How do I make a state file require another state file?
08:08 crocket Is it even possible?
08:11 LotR http://docs.saltstack.com/en/latest/ref/states/requisites.html#require-an-entire-sls-file
08:11 crocket That's a state requiring an entire sls file.
08:11 crocket I was wondering if an sls file could require another sls file.
08:12 LotR um
08:12 LotR you're going to have to explain what you think the difference between those two is
08:12 crocket I want a.sls to be executed after b.sls
08:14 kal0pr joined #salt
08:16 AndreasLutro crocket: include: [ b ] at the top of your file, then require: [ sls: b ]
08:16 AndreasLutro crocket: just like the exampel LotR gave you
08:17 crocket Does that make a state require another sls file? or does it make an entire state file require another sls file?
08:17 crocket A state file contains multiple states.
08:18 AndreasLutro try it
08:20 s_kunk joined #salt
08:22 s_kunk joined #salt
08:22 s_kunk joined #salt
08:22 kal0pr joined #salt
08:26 zer0def joined #salt
08:27 soren joined #salt
08:27 jaybocc2 joined #salt
08:35 Xevian joined #salt
08:36 N-Mi joined #salt
08:37 kal0pr joined #salt
08:43 radone joined #salt
08:47 ITChap joined #salt
08:47 ITChap Hi,
08:48 ITChap when using gitfs does the env_whitelist get interpreted after the gitfs_base setting ?
08:49 ITChap my qitfs_base is set to follow my branch "X" so in my gitfs_env_whitelist do I need to whitelist X or base ?
08:50 arount joined #salt
08:51 arount hi there, There is a way to get grains data of specific minion using in a state ? I'm assuming 'salt["grains.get"](..)' only allow to retrieve specific grain value of the current minion's grains, I'm right ?
08:51 iwishiwerearobot joined #salt
08:53 arount The thing is I have a file that list all our servers and want to update it automatically when a new server is up, so I loop in my minions using 'salt["mine.get"]' but some usefull data are stored in the grain of minion, so I need to load it in the loop
08:54 madpenguin joined #salt
08:57 N-Mi joined #salt
09:02 crocket Thanks for support.
09:02 crocket left #salt
09:03 jaybocc2 joined #salt
09:09 ndrei joined #salt
09:14 ITChap joined #salt
09:19 otter768 joined #salt
09:25 phx joined #salt
09:25 sectionm1 joined #salt
09:32 iwishiwerearobot joined #salt
09:34 lb1a joined #salt
09:49 ingslovak joined #salt
09:54 jaybocc2 joined #salt
09:55 khebbie joined #salt
09:59 zerthimon joined #salt
10:00 kal0pr joined #salt
10:17 aqua^c joined #salt
10:21 khebbie joined #salt
10:30 jaybocc2 joined #salt
10:35 AmitC joined #salt
10:36 AmitC bundler installation failing using saltstack gem state
10:38 kal0pr joined #salt
10:39 CeBe joined #salt
10:39 dopesong joined #salt
10:49 evle1 joined #salt
10:52 yomilk joined #salt
10:52 dopesong joined #salt
10:53 stoogenmeyer joined #salt
10:54 yomilk joined #salt
10:55 giantlock joined #salt
10:56 ndrei joined #salt
10:58 SubOracle joined #salt
11:01 dopesong joined #salt
11:01 AmitC here's the description https://gist.github.com/AmitChotaliya/8d983f598dac27c98e5d
11:03 VSpike I have two masters, A and B. B is also a minion, and has A and B defined as its masters
11:04 VSpike What's odd is that on B, salt-key -L lists B's key as being unaccepted. On A, it's accepted, but salt 'B' test.ping says "Minion did not return. [Not connected]"
11:04 VSpike Yet salt-call commands (pillar.items, state.highstate) work on B
11:05 VSpike How can I debug this? Is there any way to find out which master a minion is talking to, or which connections are active?
11:06 VSpike I'm not convinced that multimaster salt is the answer to what I want. If I just want some resilience, then having two masters but have minions pointed at one via a DNS CNAME might be better
11:06 VSpike Still curious to figure out what's going on here though
11:06 jaybocc2 joined #salt
11:06 CeBe1 joined #salt
11:17 VSpike Hm. The salt formula includes a minion config parameter called "random_master". I can't find this anywhere. Is this an old thing that is now gone? Or so new it's not in the docs yet?
11:17 VSpike Looks like master_type: failover is probably what I want, but the formula doesn't include that
11:20 otter768 joined #salt
11:26 impi joined #salt
11:29 VSpike Looks like random_master has replaced master_shuffle and the docs are just wrong
11:31 mrbigglesworth joined #salt
11:32 VSpike It's not just me is it? This will never work https://github.com/saltstack-formulas/salt-formula/blob/master/salt/files/minion.d/f_defaults.conf#L160
11:34 hasues joined #salt
11:35 gcfhvjbkn joined #salt
11:42 VSpike I'm confused by this. What is the default of "master_type"? http://docs.saltstack.com/en/latest/ref/configuration/minion.html#master-type
11:43 kdm__ left #salt
11:43 AndreasLutro VSpike: "str"
11:45 VSpike Hm, yes. It really *is* str
11:45 VSpike What does that stand for?
11:45 AndreasLutro VSpike: it's for determining the type of the "master" config value
11:45 AndreasLutro since it can be a string, a function name, or a list
11:46 VSpike The comments in the file suggest that it could be set to "Standard" https://github.com/saltstack/salt/blob/8f91da6d6c81e9e5f82ade7ce4eb480a9fdf189e/conf/minion#L28
11:46 mrbigglesworth joined #salt
11:46 VSpike AndreasLutro: I don't think that's right. You can have a list but not use failover mode. The standard mode just tries to connect to them in order and picks the first one and sticks with it, doesn't it?
11:47 VSpike Unless random_master is set, where the same applies but with pre-shuffling of the list
11:47 AndreasLutro https://github.com/saltstack/salt/blob/2015.5/salt/config.py#L300
11:47 AndreasLutro https://github.com/saltstack/salt/blob/2015.8/salt/config.py#L695
11:48 AndreasLutro looks like they're being changed in the next release
11:48 VSpike It's confusing because http://docs.saltstack.com/en/latest/topics/tutorials/multimaster.html doesn't mention these settings at all. And http://docs.saltstack.com/en/latest/ref/configuration/minion.html#master-type just says it can be "func" or "failover" but doesn't mention the other option
11:48 AndreasLutro could be a deprecated option
11:49 yomilk joined #salt
11:50 ndrei joined #salt
11:50 VSpike So what's the behavior in the current release?
11:52 AndreasLutro VSpike: https://github.com/saltstack/salt/blob/2015.5/salt/minion.py#L756-L774
11:53 leszq joined #salt
11:55 AndreasLutro I know a lot of the master/minion interaction code was rewritten to use tornado, this is probably related to that
11:55 VSpike That is a twisty piece of code :/
11:55 AndreasLutro I've seen worse!
11:56 stephanbuys1 joined #salt
11:57 AndreasLutro I don't really know enough about the master/minion interaction to give solid answers
11:57 VSpike Strange. So I have a list, but I haven't set master_type to anything (going by the multi-master tutorial). So what will it do? I can't work it out from that code
11:58 aqua^c joined #salt
12:00 stoogenmeyer hey guys, in my state files I have docker.pulled clauses, but I've noticed that if the image is already pulled then they won't try to pull a newer image if it exists. How do I make it always try to find the newest image?
12:00 stoogenmeyer nvm silly me, add force flag
12:01 ljt joined #salt
12:01 JDiPierro joined #salt
12:03 moderation joined #salt
12:04 giantlock joined #salt
12:05 AndreasLutro VSpike: I think it'll just fail, because the default option is "standard" (soon to be "str") which assumes the master option is a string
12:05 AndreasLutro fail as in a python typeerror will appear somewhere
12:05 DammitJim joined #salt
12:06 VSpike I haven't seen one. It always seems to connect to the second master, but that's anecdotal
12:09 GothAck joined #salt
12:10 AndreasLutro must be some instance check somewhere I'm not spotting then
12:12 CeBe1 joined #salt
12:14 tmclaugh[work] joined #salt
12:15 amcorreia joined #salt
12:15 jaybocc2 joined #salt
12:16 mrbigglesworth joined #salt
12:16 JDiPierro joined #salt
12:17 wiqd joined #salt
12:17 OliverMT joined #salt
12:17 ndrei joined #salt
12:18 dopesong_ joined #salt
12:22 kal0pr joined #salt
12:25 khebbie joined #salt
12:27 kawa2014 joined #salt
12:31 mrbigglesworth joined #salt
12:32 elfixit joined #salt
12:36 yomilk_ joined #salt
12:37 yomilk_ joined #salt
12:46 elfixit joined #salt
12:47 bfoxwell joined #salt
12:50 bbhoss joined #salt
12:50 impi joined #salt
12:50 fivmo joined #salt
12:51 jaybocc2 joined #salt
12:54 JDiPierro joined #salt
12:55 kawa2014 joined #salt
12:56 GothAck joined #salt
12:58 sgargan joined #salt
12:59 aqua^c joined #salt
13:00 kal0pr joined #salt
13:01 mrbigglesworth joined #salt
13:02 Tecnico1931 joined #salt
13:02 numkem joined #salt
13:08 subsignal joined #salt
13:08 furrowedbrow joined #salt
13:09 subsigna_ joined #salt
13:15 kal0pr joined #salt
13:16 hasues joined #salt
13:17 jdesilet joined #salt
13:18 icflournoy joined #salt
13:18 jerematic joined #salt
13:20 kawa2014 joined #salt
13:21 otter768 joined #salt
13:26 GothAck joined #salt
13:27 jaybocc2 joined #salt
13:30 peters-tx joined #salt
13:30 khebbie joined #salt
13:30 sgargan joined #salt
13:31 Deevolution joined #salt
13:31 mrbigglesworth joined #salt
13:34 ALLmightySPIFF joined #salt
13:34 timoguin joined #salt
13:35 simonmcc joined #salt
13:47 murrdoc joined #salt
13:49 madpenguin joined #salt
13:50 gazarsgo joined #salt
13:57 andrew_v joined #salt
13:58 sarerto joined #salt
14:00 dyasny joined #salt
14:01 mrbigglesworth joined #salt
14:04 sarerto Hi peoples. I am trying to set static routes for a number of fedora 22 machines. The problem is I do not know exactly how I am supposed to figure out each machines proper interface name. I have their IPs and the interfaces MAC address. But the active interface name differs from machine to machine.
14:04 sarerto salt.states.network.routes requires a name argument that is the interface name.
14:05 evilrob joined #salt
14:07 denys joined #salt
14:07 napsterX joined #salt
14:08 jaybocc2 joined #salt
14:10 CeBe1 joined #salt
14:12 hasues left #salt
14:15 AndreasLutro sarerto: run `salt minion-name grains.items`, I'm pretty sure network interface info is in there
14:16 streetmapp joined #salt
14:17 spark_ joined #salt
14:17 hasues joined #salt
14:17 streetmapp hi folks, having some issues with salt-ssh and sudo access with requiretty enabled. Opened up this issue that should contain the information: https://github.com/saltstack/salt/issues/25134 Any thoughts on what could be going on here? Or need more information?
14:17 hasues left #salt
14:18 leszq joined #salt
14:18 manfred streetmapp:  can you edit that and do the code indents on the lines with your commands so that it doesn't line wrap?
14:19 manfred streetmapp:  https://guides.github.com/features/mastering-markdown/#GitHub-flavored-markdown
14:20 JDiPierro joined #salt
14:21 sroegner joined #salt
14:21 sarerto AndreasLutro, network interface info is in there. But there is no way for me to find the correct interface name based on the information I have. Every machine has multiple interfaces (for example eno1, lo, vibr0). There is no way for me to select the interface that belongs to IP address 192.168.x.x
14:22 streetmapp manfred: all set. my mistake on that. i attempted to do that but was putting the quotes on the line, so was messing things up. should look better now.
14:23 subsignal joined #salt
14:23 manfred cool, unfortunately i have no idea what is wrong :/
14:27 pullphinger joined #salt
14:27 chiui joined #salt
14:31 mrbigglesworth joined #salt
14:32 lothiraldan joined #salt
14:35 IanV0rn2341 joined #salt
14:36 sgargan joined #salt
14:36 sroegner joined #salt
14:41 quantumriff joined #salt
14:42 clintberry joined #salt
14:43 timoguin joined #salt
14:45 otter768 joined #salt
14:46 mrbigglesworth joined #salt
14:47 aqua^c joined #salt
14:54 hasues joined #salt
14:54 hasues left #salt
14:58 malinoff joined #salt
14:58 quasiben joined #salt
15:01 sgargan joined #salt
15:05 JDiPierro joined #salt
15:05 codekobe joined #salt
15:05 Gareth morning morning
15:06 evle1 joined #salt
15:07 salty joined #salt
15:09 quasiben joined #salt
15:09 spark_ joined #salt
15:11 murrdoc o/
15:13 * Gareth waves at murrdoc
15:14 gnord joined #salt
15:15 yudao joined #salt
15:16 mrbigglesworth joined #salt
15:21 SheetiS joined #salt
15:21 iggy I'm officially back from vacation
15:21 iggy not too happy about that fact
15:22 tgmpje joined #salt
15:22 yudao left #salt
15:24 murrdoc :)
15:24 murrdoc o/
15:25 khebbie joined #salt
15:28 yudao_ joined #salt
15:28 linjan joined #salt
15:28 yudao_ left #salt
15:29 JDiPierro joined #salt
15:29 Guest89 joined #salt
15:30 JDiPierro joined #salt
15:31 blakev joined #salt
15:32 LyndsySimon joined #salt
15:32 zerthimon joined #salt
15:38 leszq joined #salt
15:39 tgmpje left #salt
15:44 Ahlee anybody had any luck with default returners on 2015.5.2? Neither mongo_future_return nor redis appear functional, wondering if it's the lazy loader, returners __init__, or me. Most likely me, of course.
15:44 JoshuaX joined #salt
15:46 pcdummy iggy:  :)
15:46 mpanetta joined #salt
15:46 mrbigglesworth joined #salt
15:50 pppingme joined #salt
15:52 iggy Ahlee: did you check bugs? I think I saw one about redis
15:53 murrdoc redis returner is necessary
15:53 murrdoc i mean any is necessary
15:53 Ahlee i'm setting up memcache now to test my theory of return/__init__.py
15:54 Ahlee you guys'll love this. Rather than us fix our using env instead of saltenv, i've been tasked with writing a new api that translates env into saltenv
15:54 Ahlee so i'll terminate the connection and re-establish
15:54 Ahlee shootmenowplz
15:56 iggy you guys must make pretty heavy use of saltenv... would love to read about that sometime
15:56 iggy we avoid env's like the f'ing plague
15:56 quasiben joined #salt
15:56 Ahlee I wish we had.
15:56 Ahlee It's all still fallout from our failed gitfs
15:57 Ahlee it's just all our external calls to the api all reference an env=
15:57 kal0pr joined #salt
15:57 murrdoc nginx rewrite ?
15:57 Ahlee so it'd be easy to pull it from salt, not so easy to update the few hundred remote calls
15:57 iggy what murrdoc said
15:57 _JZ_ joined #salt
15:57 iggy he says smart things every once in a while
15:57 murrdoc \o/
15:58 Ahlee murrdoc: hrm. that just shot up to be my afternoon project
16:00 Gareth Ahlee: specifics of the returners not functioning?
16:00 murrdoc dont get loaded in
16:00 murrdoc also redis returner doesnt purge old jobs
16:01 murrdoc not a fan
16:02 Ahlee Gareth: https://gist.github.com/jalons/46f07c50385c2a283e8f
16:02 Ahlee thinking this is jsut my bad luck
16:03 Ahlee doesn't look like mongo is actually populating mongo.db, that's coming through as '', so trying to connect to localhost
16:03 Ahlee redis i haven't dove into in depth yet
16:03 Ahlee trying memcache now just to verify
16:03 Gareth that bottom one.  I swore I fixed that.
16:04 murrdoc also with redis you have to specify the port, default doesnt work
16:04 Ahlee Gareth: v2015.5.2 tag
16:04 Ahlee murrdoc: redis.db, redis.host, redis.port defined
16:04 Gareth Ahlee: yeah.  Almost certain the fix was against 2015.5
16:04 * Gareth looks
16:04 Ahlee whoops, i leaked hostnames
16:05 khebbie joined #salt
16:05 * Gareth hacks Ahlee
16:05 * iggy hacks
16:06 murrdoc 2014.7 isnt supported anymore
16:06 sgargan joined #salt
16:06 Gareth Ahlee: back in May: https://github.com/saltstack/salt/pull/24251
16:06 * murrdoc adds todo list item
16:07 andygabby joined #salt
16:08 jeremati_ joined #salt
16:08 Ahlee Gareth: interesting. I show it as present in my branch too
16:09 Ahlee i'll do some more poking
16:10 Ahlee as why should that have come through as None
16:12 stoogenmeyer joined #salt
16:16 mrbigglesworth joined #salt
16:16 S_Walker joined #salt
16:17 stanchan joined #salt
16:20 sgargan joined #salt
16:21 Bryson joined #salt
16:23 pcdummy joined #salt
16:24 pcdummy joined #salt
16:24 pcdummy joined #salt
16:24 jaybocc2 joined #salt
16:26 conan_the_destro joined #salt
16:26 rdas joined #salt
16:27 _JZ__ joined #salt
16:27 aparsons joined #salt
16:34 murrdoc joined #salt
16:34 berserkir joined #salt
16:35 aparsons joined #salt
16:36 aqua^c joined #salt
16:37 amcorreia joined #salt
16:38 jerematic joined #salt
16:39 fresco joined #salt
16:39 fresco Hi, I have question that I really could use help with
16:40 fresco Since the merge of salt-api into salt-master, how do I start the salt-api service?
16:42 fresco Note: cherry py is installed on my machine
16:42 geekatcmu If only cherry py were installed in my tummy
16:43 spark_ joined #salt
16:43 fresco :O
16:44 iggy fresco: most distros split the salt-api init script into it's own package (even though all the modules/etc are installed by the main salt package)
16:44 Ahlee fresco: https://github.com/saltstack/salt/blob/develop/scripts/salt-api
16:45 fresco iggy: how do I go about running it alongside the master?
16:46 Ahlee it's still a seperate daemon
16:46 mrbigglesworth joined #salt
16:46 murrdoc do you Ahlee, iggy have a reactor that handles key acceptance
16:47 murrdoc i have a simple one that i want to handle key accept, refresh all modules, sync_all and then run one state
16:47 Ahlee murrdoc: open_mode here
16:47 murrdoc u lucky lucky man
16:47 iggy same
16:48 iggy but we have a reactor that does sync_all
16:48 iggy it's about what you'd expect
16:48 murrdoc what do u link it to
16:48 murrdoc like what tag
16:49 murrdoc - 'salt/auth':
16:49 murrdoc - /srv/salt/state/global/reactor/saltkey.sls
16:49 murrdoc and then in the reactor what i have
16:49 murrdoc {% if 'act' in data and data['act'] == 'pend' %}
16:49 murrdoc {% endif %}
16:49 murrdoc 2014.7
16:49 iggy - 'salt/minion/*/start':
16:49 iggy - salt://salt/reactor/sync_all.sls
16:50 fresco iggy: i checked all the running daemons and the salt-api is not running
16:50 murrdoc yeah cant do that
16:50 murrdoc salt needs to add more sensible events
16:50 murrdoc like a first run event would be so aewsome
16:50 iggy and then although, maybe I should be using auth (I think the auth signal still gets sent in open_mode)
16:50 ingslovak joined #salt
16:51 synical joined #salt
16:51 iggy murrdoc: I'd +1 a PR for that
16:51 CeBe1 joined #salt
16:51 iggy it seems like a lot of people have had to solve that particular problem one way or another
16:51 iggy (and "onboot")
16:51 murrdoc Gareth:  how hard would it be to add a 'First Run' event
16:51 murrdoc and an 'On Boot'
16:52 murrdoc shit i ll write the code, someone just point me at it where i can do it from
16:52 davedash joined #salt
16:52 pcn Has anyone used the jinja2 select filter to determine if e.g. a directory has a prefix match?
16:52 pcn I'm not clear if it's possible because the examples are not helpful in the jinja docs
16:53 iggy pcn: {% if foovar.startswith('someprefix') %} .... {% endif %}
16:54 MrMorden joined #salt
16:54 beauby joined #salt
16:55 pcn That's better than trying to use jinja filters?
16:55 pcn I was hoping to make the template more concise
16:55 leszq joined #salt
16:56 GothAck joined #salt
16:56 jbub joined #salt
16:56 Gareth murrdoc: first run event?
16:56 Gareth murrdoc: scheduled jobs?
16:57 murrdoc nah reactors
16:57 murrdoc just checking if u had any info
16:57 murrdoc info/insight
16:58 Gareth Ahh.
16:58 iggy pcn: filters in this context don't make a whole lot of sense, can you try to be more precise?
16:58 Gareth murrdoc: wouldn't think it would be too hard.  pretty sure the minions already fire an event back to the master when they start up and connect.
16:59 murrdoc Gareth:  they fire the SAME event each restart of the minion
16:59 iggy the issue is only firing that once (for the special case of a machine coming up once)
17:00 murrdoc yeah
17:00 Ahrotahntee does anyone here use salt-minion on a windows server, and would be willing to give a brief summary (or reference link) to how salt configures something like IIS?
17:00 Gareth ah in other words, event firing based on minion starting AND the amount of time the machine has been up.
17:00 murrdoc yeah
17:01 murrdoc if uptime < 5 minutes .. fire back-from-reboot event
17:01 iggy it's not unsolvable
17:02 Gareth I wonder if you could set a grain on boot somehow, look for the presence and value of that grain when determining if you need to fire the event or not.
17:02 iggy but it seems like quite a few people have gone down the path and ended up with a number of different ways
17:02 iggy so a canonical (or better yet built-in) way would be best
17:02 Gareth the trick would be clearing the grain between reboots, shutdowns.
17:07 murrdoc i dont think it needs to be a grain per se
17:08 Ahlee ext_pillar it
17:08 jbub joined #salt
17:08 Ahlee 'cause, you know, state is so awesome
17:08 iggy set a grain with last boot time and if now-uptime > last boot time, then there was a reboot? (or something like that)
17:08 murrdoc i mean … if uptime < 5 minutes is a safe enough thing
17:08 murrdoc why add more to it ?
17:09 keimlink joined #salt
17:11 iggy I'm more thinking along the lines of "if enough people are solving this problem, it makes sense for it to be built-in to salt
17:12 murrdoc yeah and / or point me at what other people did so there is re use
17:12 murrdoc i was hoping beacons would have an on boot
17:12 murrdoc but they dont
17:15 theologian joined #salt
17:15 viq joined #salt
17:16 mrbigglesworth joined #salt
17:18 aw110f joined #salt
17:21 soren joined #salt
17:22 spark_ joined #salt
17:23 spookah joined #salt
17:24 Tecnico1931 joined #salt
17:27 madpenguin joined #salt
17:28 impi joined #salt
17:35 aron_kexp joined #salt
17:36 lexter joined #salt
17:39 denys joined #salt
17:40 venu0336 joined #salt
17:42 Gareth murrdoc: wouldn't be too hard to write a beacon to monitor it.
17:43 Gareth I might start hacking on one later.
17:44 stoogenmeyer joined #salt
17:44 Ahlee memcache returner also failing. Should have specified, this is client side i'm seeing gareth
17:45 Ahlee so, tracking that down now
17:45 Gareth Ahlee: seems related to that gist you posted with the exception.
17:45 Ahlee Gareth: indeed.
17:45 Ahlee [DEBUG   ] memcache server: :
17:45 Ahlee "Well there's your problem right there"
17:46 Gareth any luck finding the cause of that exception?
17:46 Ahlee n'yet
17:46 Ahlee had to step away to do the part of my job that pays me
17:46 Gareth what does the memcache bits of the minion config look like?
17:46 Ahlee stupid work, expecting me to work
17:46 mrbigglesworth joined #salt
17:47 sovern joined #salt
17:47 sovern I installed salt on centos 7 from pip, and keep getting module missing errors.
17:47 sovern Does pip not install all of salt's dependencies?
17:47 Ahlee Gareth: have tried the following (with both master and minion being brought down between runs): https://gist.github.com/jalons/894f92766a3806da1b7c
17:48 Ahlee Gareth: have tried DNS values and IPs for all my tests, mongo, redis, now memcache
17:49 Gareth sovern: it doesn't install all the dependancies that every module needs.
17:49 Ahlee so, i'm looking through _fetch_option in returners/__init__.py to trace through this all
17:49 sovern Gareth: is there a way to get all the dependencies installed?
17:50 Gareth short of grepping through all the imports, figuring the packages and installing them.  I don't think so.  Which ones are missing for you?
17:51 Gareth Ahlee: returner is being called from a scheduled job or from the CLI?
17:51 KyleG joined #salt
17:51 KyleG joined #salt
17:51 aparsons joined #salt
17:52 iggy sovern: the problem becomes... not everybody needs every module... so you could potentially end up with a metric shit ton of unnecessary modules installed
17:52 Ahlee Gareth: to my knowledge, ext_job_cache being set appends that value as --return to all calls
17:53 Ahlee Gareth: so, I'm not specifically adding any, rather letting it be appended on it's own. I'll disable the ext_job_cache setting for now and try with just --return
17:54 dyasny joined #salt
17:54 Gareth Ahlee: I haven't used ext_job_cache, was just curious how it was being called since the logic in the __init__.py for grabbing the configs is different depending on how it's called.
17:54 Ahlee Gareth: thanks for the heads up. I'll take a step back and start from the beginning
17:54 Gareth I have to run off in a bit but let me see if I can pin point the area where the problem is liking happening.
17:55 Ahlee no worries, this has become my day
17:56 * iggy fetches bandages for Ahlee's wrists
17:56 * murrdoc fetches alcohol
17:56 Ahlee it's in _feetch_option (__init__.py:117), cfg is coming through as not a dictionary, so c_cfg isn't being populated
17:56 Gareth Ahlee: do you still have that gist with the exception?
17:57 Ahlee Gareth: i'll put it back up, deleted due to hostnames
17:57 Gareth k
17:58 giantlock joined #salt
17:58 throwawayj joined #salt
17:58 prwilson_ joined #salt
18:00 catpig joined #salt
18:00 throwawayj i have setup install of ncpa-1.8.1.exe using the salt win repo.  It installs fine but when i run pkg.list_pkgs the NCPA app doesn't show up. logging into the server i installed ncpa doesn't show the program listed under installed programs. Any idea why this may be?
18:00 yomilk joined #salt
18:03 Ahlee Gareth: https://gist.github.com/jalons/67f68ba52f2fd6373492
18:07 clintber_ joined #salt
18:08 khebbie joined #salt
18:10 Gareth Ahlee: have you tried making the config looking like this: https://gist.github.com/garethgreenaway/8cc6d48c39cac4065ccf
18:11 otter768 joined #salt
18:11 bastiandg joined #salt
18:12 Ahlee Gareth: I have not. I shall try that
18:12 notnotpeter joined #salt
18:14 Gareth Gotta run. bbl.
18:14 Ahlee Gareth: thankfully for me, no change
18:14 Ahlee i'll keep you posted
18:15 throwawayj any one here worked with using salts win repo?
18:16 mrbigglesworth joined #salt
18:18 quique joined #salt
18:20 X67r joined #salt
18:21 joeto joined #salt
18:21 icflournoy joined #salt
18:23 ndrei joined #salt
18:23 linjan joined #salt
18:25 quique I have this snippet in my state: http://fpaste.org/240211/ i have a reactor that is watching for the ebsattach tag: http://fpaste.org/240215/ that kicks off a runner that attaches the ebs volume.  I would like the next state (ebsmount) to wait until the runner finishes and returns before it tries to mount the ebs volume.  Is there a way to do that?
18:25 aqua^c joined #salt
18:25 sgargan joined #salt
18:26 iggy quique: fire a custom event at the end of the state that runs after ebsattach then trigger ebsmount from that?
18:26 aw110f joined #salt
18:27 quique iggy: fire the event from with in the runner?
18:27 iggy from wherever makes the most sense
18:28 quique ok yeah that might work, i'll try that thanks
18:28 iggy I'm too lazy to open a bunch of browser tabs to see what you're actually doing, but that's the general idea
18:31 baweaver joined #salt
18:31 Bryson joined #salt
18:32 aparsons joined #salt
18:42 johnkeates joined #salt
18:46 mrbigglesworth joined #salt
18:47 dopesong joined #salt
18:49 joehh joined #salt
18:54 baweaver joined #salt
18:56 quasiben joined #salt
18:57 hybridpollo joined #salt
18:58 julez joined #salt
18:59 ajw0100 joined #salt
19:00 sunkist joined #salt
19:05 CeBe1 joined #salt
19:05 SubOracle joined #salt
19:07 LV-426 joined #salt
19:11 fllr joined #salt
19:14 dustywusty Does anyone know if you can specify an env for a local.cmd.run in a reactor state?
19:15 tvinson are multi-line keys possible in salt yaml? i've got a pillar/top.sls that looks something like http://pastebin.com/eZgbrAJc i'd like to be able to target a list of hosts for a role but it's highly desirable to avoid big long unreadable lines.
19:15 tvinson multi-line keys don't have to be the right answer if someone has a better way to do this
19:16 mrbigglesworth joined #salt
19:18 losh joined #salt
19:18 jalbretsen joined #salt
19:18 iggy tvinson: 99% sure the answer is no
19:20 sunkist joined #salt
19:20 ipmb joined #salt
19:26 baweaver joined #salt
19:27 JDiPierro joined #salt
19:27 baweaver joined #salt
19:28 kula tvinson: there's a yaml way of specifying large keys, I don't remember what it is off hand, but I've used it to identify ssh keys in the past.
19:30 gcfhvjbkn joined #salt
19:36 VSpike tvinson: yep
19:36 VSpike https://github.com/saltstack-formulas/salt-formula/blob/master/pillar.example#L119
19:36 VSpike tvinson: ^
19:36 VSpike example
19:36 s_kunk joined #salt
19:36 s_kunk joined #salt
19:36 VSpike the vertical bar starts a multiline string in YAML
19:37 baweaver joined #salt
19:37 quasiben joined #salt
19:37 tvinson VSpike: right, but that's for a value. i'm looking for multi-line keys. they're possible in yaml but I think the linebreaks map to spaces which breaks my list match type.
19:38 N-Mi joined #salt
19:38 murrdoc joined #salt
19:38 murrdoc how bad is it
19:38 sastorsl left #salt
19:38 murrdoc to have a map.jinja in your top.sls
19:38 JDiPierro joined #salt
19:39 VSpike Oh, sorry... misunderstood. Read keys as SSH keys :)
19:39 tvinson VSpike: no worries, i can see how you would read it that way
19:43 quantumriff this might seem silly.. but when I do a salt -v '*' state.highstate.. is there a way to sort the minions alphebetically?
19:43 btorch_ is this some salt bug ?
19:43 btorch_ http://pastebin.com/Pbje4ALT
19:45 btorch_ seems to only happen when I use -b
19:46 mrbigglesworth joined #salt
19:48 cberndt joined #salt
19:49 iggy search for existing bugs and if none exist, file a new one
19:49 pcn Does salt-cloud provide any grains from e.g. ec2 for instance-id, etc?
19:49 iggy the devs generally don't like tracebacks
19:49 iggy pcn: salt-contrib has some
19:49 gcfhvjbkn joined #salt
19:56 quique iggy: i made a reactor in the master config that calls an sls file that does the mount but it errors: Exception caught by reactor: cmd_async() takes at least 3 arguments (2 given)
19:59 spark_ joined #salt
19:59 quique how does the reactor know what minion to execute it on?
19:59 quique I'm guessing that is probably the issue
20:02 iggy quique: the docs show how to set the target to the minion that sent the event
20:03 iggy data['id'] I think
20:05 quique i'm sending the event from the master
20:05 iggy oh right
20:06 iggy why are you doing this stuff from a runner again btw?
20:06 iggy (vs just having the reactor run a state on the minion to do the attach)
20:06 quique i want the script to run on the master
20:06 quique so that aws creds aren't sent to the minion
20:09 iggy fair enough, I guess you'll have to figure out some way to pass the minion with the event (no clue how to do that)
20:14 aqua^c joined #salt
20:16 soren joined #salt
20:16 mrbigglesworth joined #salt
20:19 baweaver joined #salt
20:23 quasiben joined #salt
20:25 zmalone joined #salt
20:31 giantlock joined #salt
20:32 sgargan joined #salt
20:32 markm joined #salt
20:33 blakev_ joined #salt
20:35 ndrei joined #salt
20:39 gcfhvjbkn joined #salt
20:44 aparsons joined #salt
20:46 mrbigglesworth joined #salt
20:49 ndrei joined #salt
20:58 quique joined #salt
21:03 baweaver joined #salt
21:05 madpenguin joined #salt
21:08 subsigna_ joined #salt
21:08 quique iggy: am i missing something but it looks like there's no way for runners to be inserted in an order dependent collection of states.  for example here is my top.sls: http://fpaste.org/240270/.  The ebsattach state in in common.  it fires off the event and returns right away that it kicked it off successfully and continues to all the other states, but the other states depend on the ebs volume being mounted.  I am just thinking about this all wrong?
21:10 quique even if i figure out how to pass on the minion id with the event, i don't think i have a way to insert it back into the domino of the states
21:10 iggy quique: I don't think you can run runners from a highstate if that's what you're asking
21:11 Ahlee cmd.run a call out to the api
21:11 Ahlee lol.
21:11 Ahlee (Please don't do that)
21:11 * iggy shoots first, asks questions later
21:12 iggy you could use orchestrate
21:13 whytewolf humm. just coming into this conversation and i have to toss in my two cents orchenstrate would most likely be a clearner option.
21:13 Ahlee can you summarize what you're trying to do quique? I don't want to scroll back and catch up
21:14 Ahlee "no" is of course acceptable answer
21:14 throwawayj exit
21:16 Ahlee foo = inspect.getsource(cfg)
21:16 Ahlee er
21:16 mrbigglesworth joined #salt
21:16 Ahlee one day i'm going to learn tmux copy/paste buffers.
21:16 quique minion comes up: if it has an ebs_volumes grain, i want to (from the master to keep aws creds off minion) create and attach an ebs volume, then mount it, and continue with the rest of the states
21:17 quique Ahlee: but all the subsequent stuff needs the volume there and mounted
21:17 Ahlee right, so you need to block until aws spins up and makes the volume available
21:17 Viper233 joined #salt
21:17 Ahlee which could be n time
21:18 Ahlee orchestrate i guess
21:18 Ahlee as others have said
21:18 quique i'll look into that
21:18 Ahlee or just push the creds to the minion and YOLO
21:18 quique ?
21:18 quique lol
21:18 keimlink joined #salt
21:18 Ahlee you siad you don't want to push the credentials to the minion. the possibility exists to just do it there, as it'd be easier
21:19 quique Ahlee: what is the gist of orchestrate?
21:19 iggy I still think events are the way to go
21:19 iggy I just wouldn't use a runner
21:19 quique iggy: what would be a better way?
21:20 sgargan joined #salt
21:20 iggy use a normal state that's only targeted at the master to do the ebs stuff, use a dbus beacon on the minion that fires off an event when the minion actually sees the disk, BAM!
21:21 sectionme joined #salt
21:21 iggy there's a journald beacon already written
21:22 Ahlee unrelated, beacons run in a seperate python interpretor than the master right?
21:23 iggy a dbus beacon isn't a bad idea in general (would work on more than just aws), but I don't think one exists yet
21:23 iggy Ahlee: there was discussion of that at the conf... I don't remember what they said (but something tells me no)
21:24 Ahlee yeah, i do'nt remember which was decided, either
21:24 Ahlee i can argue both ways
21:24 Ahlee given my cronjob that cleans up python core files on the master, I know which way i'm hoping.
21:24 iggy You're doing it wrong
21:25 Ahlee ain't that the truth.
21:25 Ahlee so https://github.com/saltstack/salt/blob/develop/salt/modules/config.py#L129 is my issue with returners
21:28 cberndt joined #salt
21:33 quique whytewolf, Ahlee so looking at orchestrate is interesting, but how would i do it? is there a way to run something on the master as part of the chain of states?
21:33 Ahlee last updated 1/14/13...doubt this is my issue.
21:34 ajw0100 joined #salt
21:35 sunkist joined #salt
21:35 Ahlee quique: you'll need a way to target your master
21:36 Ahlee i.e. set a grain on the master to be saltmaster: True
21:36 Ahlee or something
21:36 soren joined #salt
21:36 quique ok, i see
21:36 quique thanks
21:37 baweaver joined #salt
21:39 anotherZero joined #salt
21:45 iggy oh cool, there is already a _beacons
21:45 aparsons joined #salt
21:50 dark_magic1339 joined #salt
21:52 baweaver joined #salt
21:54 mrbigglesworth joined #salt
21:55 baweaver joined #salt
21:55 dopesong joined #salt
21:56 mrbigglesworth joined #salt
21:59 linjan joined #salt
22:02 aqua^c joined #salt
22:11 cwyse joined #salt
22:12 nethershaw joined #salt
22:19 subsignal joined #salt
22:21 troyready joined #salt
22:24 wenzel_ joined #salt
22:25 wenzel_ hi !
22:25 wenzel_ is it possible to use salt-ssh with an alternative root ?
22:27 iggy -c
22:28 wenzel_ this will replace /etc/salt with my local path
22:28 wenzel_ but how to replace /var/cache/salt and /src/salt ?
22:28 iggy in the config file...
22:29 wenzel_ but these would be absolute paths
22:30 wenzel_ is there a way to pass both cache dir and root dir as command line parameter ?
22:30 kaictl joined #salt
22:31 iggy if salt starts reading/writing to the cache dir before it's read the cachedir option from the config file, that'd be a bug that you should file
22:32 anotherZero joined #salt
22:40 wedgie i've installed an alternate version of python and pip on a minion. Is there a way to amke the pip_state work with the alternate pip install? Centos comes with python 2.6 by default but i've installed 2.7 and pip2.7
22:41 iggy start salt with the other python
22:49 quasiben joined #salt
22:49 ajw0100 joined #salt
22:50 JDiPierro joined #salt
22:56 JDiPierro joined #salt
22:57 dopesong joined #salt
23:02 fxhp joined #salt
23:03 mosen joined #salt
23:10 sectionme joined #salt
23:15 bezeee joined #salt
23:20 hybridpollo joined #salt
23:21 dimeshake joined #salt
23:22 baweaver joined #salt
23:23 c10 joined #salt
23:34 ndrei joined #salt
23:34 subsignal joined #salt
23:37 jeddi joined #salt
23:38 subsignal joined #salt
23:40 jerematic joined #salt
23:40 yomilk joined #salt
23:45 tmclaugh[work] joined #salt
23:50 KyleG1 joined #salt
23:51 aqua^c joined #salt
23:53 joehh basepi: just wondering about windows releases
23:53 joehh http://docs.saltstack.com/downloads/ has 2015.5.3, but as far as I am aware only 2015.5.2 has been released
23:54 joehh but the docs only link to 2015.5.0 at http://docs.saltstack.com/en/latest/topics/installation/windows.html
23:54 joehh is 2015.5.2 the right package to use?
23:55 joehh I'm sure the answer is "test in your environment", but some clarity on the differences between 2015.5.2 and the others would be great
23:58 Gareth Ahlee: any luck?
23:58 sgargan joined #salt

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