Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-03-03

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

All times shown according to UTC.

Time Nick Message
00:08 rgbkrk joined #salt
00:16 Math` joined #salt
00:27 schimmy joined #salt
00:30 mgw joined #salt
00:35 Math` joined #salt
00:36 faldridge joined #salt
00:39 srage joined #salt
00:43 xmltok joined #salt
00:45 justlooks joined #salt
00:47 diegows joined #salt
00:48 viod joined #salt
00:51 sdelisle joined #salt
00:52 sdelisle Hello, I'm having some trouble with state pkg.installed
00:53 sdelisle 'Percona-Server-server-51' not found (possible matches: Percona-Server-server-51)
00:54 sdelisle Anyone have an idea what really might be missing the first time around?
00:54 jeffro sdelisle: sure it's not an apt problem?
00:54 jeffro i think with percona stuff you need to apt-add source their repo first
00:54 sdelisle yum, but no I'm not sure
00:54 jeffro ah yum
00:55 EWDurbin "Comment: Package 'postgresql93-server' not found (possible matches: postgresql93-server)"
00:55 jeffro do you have a percona-server-server-51 package declared somewhere in an sls?
00:58 alienresidents joined #salt
01:00 krow joined #salt
01:01 sdelisle joined #salt
01:02 beach from an architecture point of view... perhaps security more than anything. Are most folks keeping their cloud providers locally and not putting them in gitfs?
01:04 sdelisle jeffro: yes I have states for a bunch of packages and they all get this error on the first salt run but a scond run is fine
01:05 sdelisle I'm curious about the comment about postgresql ... is someone else experiencing this problem?
01:08 faldridge joined #salt
01:08 Hefeweizen joined #salt
01:09 jeffro sdelisle: my guess is you're forgetting to declare a dependency for one of the packages
01:09 jeffro it works the second time because its installed
01:09 jeffro but the first time it gets installed after the package that depends on it
01:09 fllr joined #salt
01:11 EWDurbin sdelisle: yeah i'm getting hit with this too
01:11 EWDurbin specifically in 2014.01 on CentOS 6
01:13 EWDurbin it appears that unwinding my config based yum repos into pkgrepo.managed again (i had gone with a file after a 0.17 release broke pkgrepo.managed...)
01:13 EWDurbin but i'm running from scratch to verify
01:18 EWDurbin nah, no good
01:18 EWDurbin still requires a second run
01:19 jeremyfelt joined #salt
01:21 malinoff joined #salt
01:22 sdelisle jeffro: that was my though too, but then I would expect an error message either from salt or from yum, but the only error is that enigmatic one from salt.
01:23 jeffro sdelisle: yeah that would be nice hehe. im pretty new to salt but im pretty sure it isn't able to know if a package has a dependency, you will have to declare it yourself i think
01:23 jeffro salt will just complain at you if you break their syntax or dependency rules
01:24 MedicalJaneParis joined #salt
01:25 sdelisle jeffro: but then yum will let you know that it can't find the dependency.  In this case salt is not getting to the point of running yum
01:25 justlooks how to replace a sentence in a file using salt?
01:25 schimmy joined #salt
01:26 malinoff http://docs.saltstack.com/ref/states/all/salt.states.file.html?highlight=states.file#salt.states.file.blockreplace
01:26 mgw joined #salt
01:26 malinoff Or http://docs.saltstack.com/ref/states/all/salt.states.file.html?highlight=states.file#salt.states.file.replace
01:26 faldridge joined #salt
01:32 NotreDev joined #salt
01:32 srage joined #salt
01:33 jeffro i see
01:34 jeffro and there's a package declaration in the sls for the dep?
01:36 sdelisle There are actually no dependencies for this package that aren't taken care of by yum.
01:38 jeffro sdelisle: hmm, i think im stumped then
01:38 jeffro definitely when i have problems like this it's because i forgot to require a dependency
01:38 jeffro but i haven't run into too many of the tough issues so far :P
01:38 justlooks malinoff: what if i do not know the replace content is already exist in file?
01:39 dave_den joined #salt
01:39 forresta joined #salt
01:39 malinoff justlooks, "pattern" argument is just a regex pattern
01:39 dren_ joined #salt
01:40 malinoff So if it is not found, salt will do nothing
01:41 justlooks malinoff:  what i want to do is the record is not exist on file ,it will append into it ,if the record exist ,do not change it
01:43 malinoff justlooks, the forehead decision is to use jinja to search for a content, if it is present, use replace, use file.append otherwise
01:43 rgbkrk joined #salt
01:47 gnugnu8 joined #salt
01:47 generj joined #salt
01:53 jeffro joined #salt
01:58 kintel joined #salt
02:01 generj so I'm trying to use salt-cloud
02:02 generj I installed it as per the docs (pip install on fresh ubuntu)
02:02 generj [WARNING ] The cloud driver, 'azure', configured under the 'my-azure-config' cloud provider alias was not loaded since 'azure.get_configured_provider()' could not be found. Removing it from the available providers list
02:02 generj so this is some weirdness that has to do with salt-cloud merging into mainline salt?
02:02 res0nat0r joined #salt
02:02 generj I'm on salt-cloud 0.8.11
02:04 Networkn3rd joined #salt
02:05 quantumriff joined #salt
02:06 NotreDev joined #salt
02:08 jeremyBass joined #salt
02:09 quantumriff joined #salt
02:11 quantumriff joined #salt
02:17 dangra joined #salt
02:21 danielbachhuber joined #salt
02:34 jankowiak joined #salt
02:43 sdelisle joined #salt
02:44 sdelisle I guess I figured out my issue with repos ... testing it now
02:44 sdelisle when I create a new repo, I have to make sure that the first salt state that uses it has refresh: True
02:45 sdelisle I don't think it used to be that way, I have been creating repos and fetching packages from them for a long time.
02:46 jeffro ahh
02:47 jeffro yeah i dont like the way i have mine set up... they always pull. dont really want that i dont think
02:49 faldridge joined #salt
02:49 baniir joined #salt
02:53 jslatts joined #salt
02:58 Math` joined #salt
03:03 schimmy joined #salt
03:06 Furao joined #salt
03:08 rma_ joined #salt
03:09 rma_ how to resolve package version conflict using saltstack for multiple machines
03:09 rma_ in redhat based Linux
03:11 dangra_ joined #salt
03:13 ravibhure joined #salt
03:13 yomilk joined #salt
03:14 srage joined #salt
03:14 Math` joined #salt
03:14 lionel joined #salt
03:17 ravibhure left #salt
03:22 rostam joined #salt
03:24 tmwsiy_ joined #salt
03:39 tmwsiy__ joined #salt
03:39 micko joined #salt
03:39 Ryan_Lane joined #salt
03:55 favadi joined #salt
04:13 Gifflen joined #salt
04:20 gjaber joined #salt
04:20 gjaber Hi
04:21 ravibhure joined #salt
04:21 gjaber I have a question regarding the pip state module
04:21 gjaber what is the best way to pass extra arguments to the pip call??
04:21 gjaber I need to pass the allow-external and allow-unverified flags to the pip call
04:29 krow joined #salt
04:30 ipalreadytaken joined #salt
04:32 gjaber I have a question regarding the pip state module
04:32 gjaber what is the best way to pass extra arguments to the pip call??
04:33 gjaber I need to pass the allow-external and allow-unverified flags to the pip call
04:33 malinoff gjaber, I think, it's better to ask on salt-users mailing list
04:34 xl1 joined #salt
04:34 Ryan_Lane joined #salt
04:46 jeremyfelt joined #salt
04:46 faldridge joined #salt
04:58 jdenning joined #salt
05:03 lionel joined #salt
05:16 fllr joined #salt
05:32 schimmy joined #salt
05:36 Ryan_Lane joined #salt
05:37 NotreDev joined #salt
05:41 Ryan_Lane joined #salt
05:43 ndrei joined #salt
05:50 Andrevan joined #salt
06:16 bhosmer joined #salt
06:18 NotreDev joined #salt
06:23 xet7 joined #salt
06:31 smcquay joined #salt
06:33 fllr joined #salt
06:34 ndrei joined #salt
06:36 Ryan_Lane1 joined #salt
06:37 fllr joined #salt
06:43 justlooks can salt modifiy minion sudo file ?
06:45 justlooks when i deploy some app i need edit minion's /etc/sudoers file
06:46 favadi joined #salt
06:46 malinoff justlooks, its better to build a native package and do any dangerous stuff in postinstall script
06:46 malinoff But yes, since salt is running as root by default, you can do whatever you want
06:48 justlooks malinoff: you mean i can use file.append to edit /etc/sudoers as   a common file?
06:48 malinoff justlooks, hm
06:48 malinoff dunno
06:48 malinoff never tried that :)
06:49 malinoff justlooks, you can "chown 700 /etc/sudoers" at first - this will give you access to write
06:50 faldridge joined #salt
06:50 malinoff But modifying sudoers is a bad practice
06:51 justlooks malinoff:  my problem is i have a monitor script need read /var/log/message ,but it run as user nagios ,when i call it from monitor box ,it log can not be read ,so it failed, and i need deploy it on many box
06:51 d10n joined #salt
06:56 malinoff justlooks, well, you can easily append lines at /etc/sudoers, as I just checked
06:58 jdenning joined #salt
06:58 n8n joined #salt
07:00 ndrei joined #salt
07:09 justlooks how can i use multi require ? i get error when i try like this https://gist.github.com/justlooks/9319838
07:12 malinoff justlooks, commented your gist
07:15 justlooks malinoff: i commented it
07:16 Furao joined #salt
07:20 malinoff justlooks, take a look on the last 3 rows
07:20 malinoff In my comment
07:21 jdenning joined #salt
07:22 gnugnu8 joined #salt
07:23 gammalget joined #salt
07:26 justlooks malinoff: thank you ,it worked
07:26 malinoff justlooks, the point is simple
07:27 malinoff For each requirement you should specify it's own namespace and id
07:37 Ryan_Lane1 joined #salt
07:41 ipalreadytaken joined #salt
07:42 ipalread_ joined #salt
07:42 schimmy joined #salt
07:43 fllr joined #salt
07:44 ndrei joined #salt
07:52 micko joined #salt
07:55 yomilk joined #salt
08:00 johtso joined #salt
08:01 Furao left #salt
08:03 MrTango joined #salt
08:04 lionel joined #salt
08:11 Kenzor joined #salt
08:13 harobed_ joined #salt
08:16 martoss joined #salt
08:16 lionel joined #salt
08:17 fllr joined #salt
08:17 martoss1 joined #salt
08:18 cast joined #salt
08:20 viod joined #salt
08:26 Ryan_Lane joined #salt
08:28 miguitas joined #salt
08:29 yomilk_ joined #salt
08:33 ml_1 joined #salt
08:36 zooz joined #salt
08:36 Ryan_Lane joined #salt
08:38 yomilk joined #salt
08:41 giantlock_ joined #salt
08:43 lionel joined #salt
08:43 fllr joined #salt
08:44 scott_w joined #salt
08:46 sroegner_ joined #salt
08:46 linjan_ joined #salt
08:49 lionel joined #salt
08:54 carlos_ joined #salt
08:57 kadel joined #salt
09:12 jrdx joined #salt
09:19 balboah joined #salt
09:22 che-arne joined #salt
09:23 slav0nic joined #salt
09:27 yota joined #salt
09:31 ipalreadytaken joined #salt
09:43 fllr joined #salt
09:59 Thumper650 joined #salt
10:00 ipalreadytaken joined #salt
10:03 Sypher_ joined #salt
10:09 CeBe joined #salt
10:11 Thumper650 joined #salt
10:12 jeddi I'm looking for some examples of generating multi-paragraph config files (think /etc/network/interfaces, with paras for eth0, eth1, ... , ethn) where the number of devices is defined by the presence of them in a pillar file for the host (eth0: \ address: ... \ netmask: ...\ )  etc.  I'm using a jinja template now but I can't work out how to loop - whether I do that in the state file where I retrieve pillar data, or in the jinja
10:12 jeddi template file (which would mean the .sls retrieves pillar data, and then passes them as variables to the jinja output)
10:17 cast haven't ran into that situation in salt [yet], though when i do i might see how using file.blockreplace goes for it
10:30 yomilk joined #salt
10:35 sroegner_ joined #salt
10:36 sroegner_ joined #salt
10:42 ChaosPsyke joined #salt
10:43 fllr joined #salt
10:49 Pate_ joined #salt
10:54 ravibhure joined #salt
10:58 Pate_ joined #salt
11:04 cperrin88 joined #salt
11:09 viod joined #salt
11:19 cperrin88 joined #salt
11:21 cperrin88 Hey, is there a way in salt to "delegate" a command to another server?
11:27 yes456 joined #salt
11:30 yes456 any update on  2014.1.0 ubuntu pkgs. http://ppa.launchpad.net/saltstack/salt/ubuntu/pool/main/s/salt/ still  shows 0.17.5
11:34 davidone joined #salt
11:43 fllr joined #salt
11:49 ipalreadytaken joined #salt
11:50 johtso joined #salt
11:51 bhosmer joined #salt
11:52 johtso What's the best alternative to waiting for the 2014.1.0 ubuntu package to hit the salt PPA? Are there any alternative PPAs?
11:53 faldridge joined #salt
12:00 viod1 joined #salt
12:05 ndrei joined #salt
12:05 elfixit joined #salt
12:10 yes456 johtso: i think no  :(
12:10 tmwsiy_ joined #salt
12:26 viod1 Hi! I'm trying to configure my users and their conf using salt, but can't find the error in my files: http://pastebin.aquilenet.fr/?593e1264ab69e091#ICyGB+nnbCG0K/ivByalqp+xpzoIsEZ0MbOe31C33r4=
12:26 viod1 The template renderer keeps telling me that a user has no attribute shell, even though it has one... I've compared my code with others found on the Internet, but could not see any significant difference. Any idea?
12:36 blee joined #salt
12:38 davidone left #salt
12:43 fllr joined #salt
12:44 flebel joined #salt
12:46 srage joined #salt
12:49 flebel joined #salt
12:56 ndrei joined #salt
13:01 gtmtech joined #salt
13:01 gtmtech Hi all
13:02 gtmtech I have a template and a variable like this "192.168.10" - i need to reverse it like this --> "10.168.192"
13:03 gtmtech I tried {% set reverse_ip = ip|split(".")|reverse()|join(",") %}   - but Jinja syntax error: no filter named 'split'
13:03 gtmtech Does anyone know how I can do this?
13:03 rgbkrk joined #salt
13:27 tamentis joined #salt
13:30 SaveTheRbtz joined #salt
13:30 ksk gtmtech: sounds like a python question to me ,)
13:31 gfa joined #salt
13:31 octarine joined #salt
13:34 Guest94409 joined #salt
13:35 delkins joined #salt
13:35 juice joined #salt
13:35 sroegner_ joined #salt
13:36 jslatts joined #salt
13:38 ipalreadytaken joined #salt
13:38 yes456 viod1:  error is correct pilar get converted into dict like structure
13:39 yes456 viod1: http://goo.gl/Br1Krf
13:41 viod1 yes456: so how should i call my attributes? I've tried user.['shell'], but it's not that either
13:41 topochan joined #salt
13:42 viod1 user['shell']* sorry
13:42 pt|Zool joined #salt
13:43 jakubek viod1: pillar['headers']['salt']
13:43 jakubek not
13:43 jakubek viod1: pillar['user']['shell']
13:43 fllr joined #salt
13:44 jakubek pillar['user']['root']['shell'] to get shell of root user
13:45 msciciel joined #salt
13:45 viod1 i'm trying to access it in a for loop actually, with {% for name, user in pillar['config_users'].iter() %}
13:46 nlb joined #salt
13:46 jeffrubic joined #salt
13:48 cron0 joined #salt
13:48 magenbrot joined #salt
13:52 ze- joined #salt
13:55 funzo joined #salt
13:55 gtmtech In answer to my previous question - the answer was {{ var1.split(".")|reverse()|join(".") }}
13:55 gtmtech seems some things you need to use python for, and other things you need to use filters for
13:56 dangra_ joined #salt
13:56 generj joined #salt
13:57 fxdgear joined #salt
13:58 Psi-Jack joined #salt
13:59 ndrei joined #salt
14:00 farra joined #salt
14:00 elfixit joined #salt
14:01 mband joined #salt
14:03 Ztyx1 joined #salt
14:04 jcsp joined #salt
14:04 Ztyx1 Hm, can a custom grains module reuse the results of other grains?
14:04 Networkn3rd joined #salt
14:05 lcfseth joined #salt
14:05 Networkn3rd joined #salt
14:06 yes456 viod1: http://snipurl.com/28ndj7x
14:07 davidone joined #salt
14:07 juicer2 joined #salt
14:07 yes456 viod1: not tested :(
14:10 Gifflen joined #salt
14:13 viod1 yes456: thanks, it works :D (well, I think so, as I have another error, but which is not related to template parsing)
14:14 racooper joined #salt
14:14 giantlock_ joined #salt
14:15 generj1 joined #salt
14:18 Gifflen joined #salt
14:22 vbabiy joined #salt
14:23 faldridge joined #salt
14:24 patrek_ joined #salt
14:26 viod1 mmmh maybe I said it too early
14:26 viod1 yes456: if show_highstate outputs "shell: user['shell']", it means it has not evaluated it, right?
14:27 timoguin joined #salt
14:27 faldridge joined #salt
14:27 backjlack joined #salt
14:27 ravibhure1 joined #salt
14:29 generj joined #salt
14:31 vejdmn joined #salt
14:31 fllr joined #salt
14:32 mpanetta joined #salt
14:33 _fllr_ joined #salt
14:33 Gifflen joined #salt
14:34 Gifflen joined #salt
14:36 Gifflen_ joined #salt
14:36 viod joined #salt
14:37 generj joined #salt
14:40 Gifflen joined #salt
14:43 babilen Hi all - I have a list of directories that I want to create and which are necessary for a lighttpd service to run. I have the following so far: http://paste.debian.net/85071/ but I get "Requisite declaration foo.lighttpd.dirs in SLS lighttpd is not formed as a single key dictionary" -- why is that?
14:43 fllr joined #salt
14:44 GradysGhost joined #salt
14:46 vejdmn joined #salt
14:48 babilen My impression is that I have to restructure the dirs.sls state so that it doesn't use the actual directory names as state id, but I am not sure how to best restructure it and if that is actually the issue.
14:49 jergerber joined #salt
14:52 generj joined #salt
14:54 favadi joined #salt
14:54 scurry joined #salt
14:55 abe_music joined #salt
14:55 kintel joined #salt
14:55 quantumriff joined #salt
14:56 elfixit joined #salt
14:56 generj joined #salt
14:57 quantumriff joined #salt
15:00 lyddonb joined #salt
15:02 rgbkrk joined #salt
15:03 sgviking joined #salt
15:03 generj joined #salt
15:05 generj joined #salt
15:06 quickdry21 joined #salt
15:08 viod1 joined #salt
15:08 tobiassjosten left #salt
15:15 tom[] joined #salt
15:16 MTecknology babilen: it looks like you want  - watch_in: \n\t - service: lighttpd   in that last state
15:17 generj joined #salt
15:23 rostam joined #salt
15:24 generj joined #salt
15:26 mgw joined #salt
15:26 ipalreadytaken joined #salt
15:28 opapo joined #salt
15:28 pdayton joined #salt
15:28 fxhp joined #salt
15:28 seapasulli joined #salt
15:29 naiyte joined #salt
15:29 zirpu left #salt
15:31 jeremyBass1 joined #salt
15:32 babilen MTecknology: So I would rather declare the reverse requirement than extend the lighttpd service? Shouldn't that be equivalent?
15:33 MTecknology extending and watching are different
15:34 bizboo joined #salt
15:35 jmccree Is there a way to do a something like a require but not generate warnings?
15:35 babilen MTecknology: Sure, but I mean "extend: \n\t lighttpd: \n\t service: \n\t watch: /var/www/foo"
15:36 babilen Shouldn't that be the same as "- watch_in: \n\t - service: lighttpd" ?
15:36 fxhp lighttpd should be a modular component
15:36 rojem joined #salt
15:36 babilen fxhp: Care to elaborate?
15:36 fxhp each application should extend lighttpd with watches, lighttpd should not need to know about each application
15:36 babilen Sorry, I am quite new and really do not have a feeling yet for what is idiomatic :)
15:37 fxhp No problem, it took a while for me to sort things out in my head, but alas salt lets you do it how you like
15:37 jmccree My use case being say I have a collectd sls, in the apache sls I want to enable the collectd apache plugin, but only if collectd sls is enabled for that minion
15:37 babilen fxhp: Yes, i completely agree. Isn't that kinda what I've been (trying) to do?
15:38 jmccree I do this now via a require: collectd.sls for the rules in the apache sls, but it generates errors.
15:38 jmccree I can ignore them fine, just wanted to see if I was missing a better way.
15:38 babilen fxhp: "foo" is the name of the service (well, website) .. And I essentially want to say "use lighttpd with these changes .." there
15:39 babilen Just trying to describe the "state how it should be" rather than "steps to achieve that state"
15:40 fxhp babilen - if the foo website changes, restart lighttpd
15:40 Mua joined #salt
15:41 fxhp I personally try to only use one direction of requisites, so I tend not to use the 'in' ones. (require_in, watch_in, etc)
15:41 TheRealBill_here joined #salt
15:41 ndrei joined #salt
15:42 babilen fxhp: yes, but how would I get rid of it here? It doesn't seem to create the directory if I extend lighttpd as shown.
15:42 fxhp jmccree - could you pastebin the states and errors?
15:42 jmccree fxhp, the error is the simple "require failed"
15:42 jmccree Which is to be expected.
15:42 babilen fxhp: It works as expected if I use watch_in -- I would prefer to only use "watch" though
15:43 fxhp babilen: you have a state to create the foo directory and files, that can pretty much run anytime
15:44 jmccree fxdgear, on cases where the apache sls is applied but not the collectd sls, it generates the expected require collectd sls failed.
15:44 generj I don't understand something
15:44 generj is Salt 2014.1.0 not out yet?
15:44 fxhp in that same sls, use an extend to have the lighttpd service watch that directory
15:44 generj the date sure seems to imply it is :P but all the install docs tell me to get version 0.18
15:44 fxhp generj - its out, are you looking for a package?
15:45 generj fxhp: I'm looking for any guide to install it on a fresh ubuntu 12.04
15:45 babilen fxhp: "that same sls" -- are you referring to dirs.sls in my example? Why does that have to be in the same sls?
15:45 jalbretsen joined #salt
15:45 bersace joined #salt
15:45 fxhp generj - currently fresh ubuntu package is still 0.17.x, the PPA maintainer has not finished building the 2014.1.0 package
15:46 forrest joined #salt
15:46 fxhp babilen - it doesn' have to be, I assumed you had an sls for the foo website
15:48 babilen fxhp: I have one sls for foo_website.master and one for foo_website.slave (loadbalanced with very minor changes) -- didn't want to repeat the directory states and therefore extracted those into dirs.sls
15:49 babilen So, but in summary "extend: lighttpd: service: watch: file:///path/to/whatever" should work and is a sensible approach?
15:49 fxhp babilen - ahh, is this static code?  Why do you want to restart lighttpd service on directory changes?
15:50 generj fxhp: I am happy to install manually
15:50 fxhp static "HTML"
15:50 generj fxhp: there are no instructions and salt-cloud does not work on 0.17.x
15:50 Furao_ joined #salt
15:51 Ztyx joined #salt
15:51 babilen fxhp: It is ... I was using "require" before, but switched to "watch" as it was frequently mentioned. All I care about right now is that the directories exist. I have not decided yet if I actually want to use watch or require, just that the directory has to be present *before* the service is started.
15:51 fxhp generj - I am using salt from PPA (0.17.5) + a manual salt-cloud install
15:52 fxhp generj: basically `pip install salt-cloud`
15:53 fxhp babilen - Ohh ...
15:53 generj fxhp: that worked out alright except salt-cloud would refuse to configure a provider
15:54 generj fxhp: [WARNING ] The cloud driver, 'azure', configured under the 'my-azure-config' cloud provider alias was not loaded since 'azure.get_configured_provider()' could not be found. Removing it from the available providers list
15:54 generj fxhp: so I figured it must be an outdated salt-cloud bug and as it was merged into salt proper I want to get salt 2014
15:54 generj but maybe I'm wrong
15:55 fxhp generj: sounds correct ...
15:55 fxhp generj: ok you might want to delete that version you just installed
15:55 forrest generj, salt-cloud is only part of main line salt in 2014.1.0+, for 0.17.5 you need to install it.
15:56 davet joined #salt
15:56 fxhp babilen - require sets dependency ordering.  watch sets dependency ordering AND alerts special states to changes, which can trigger reactions
15:57 generj forrest: that's what I'm saying, I had salt 0.17.5 and installed salt-cloud and hit a bug
15:57 generj forrest: so I thought maybe I could solve it by upgrading both and moving to salt 2014.01 wholesale
15:57 fxhp babilen: for instance we could have the lighttpd service watch the lighttpd conf file, which means the conf file needs to run first and if the conf file changes, the service will restart
15:57 forrest generj, what version of salt cloud did you install?
15:57 fxhp salt-cloud --verison
15:58 generj forrest: I don't know how to solve this "azure.get_configured_provider()" on salt 0.17.5 / salt-cloud 0.17.5
15:58 babilen fxhp: So, require is what I want here as I simply want to ensure that the dir state is being run before the service one? I do understand the differences between watch and require. :)
15:59 forrest generj, similar issue with ec2: https://github.com/saltstack/salt-cloud/issues/890
15:59 forrest from a few months ago
15:59 generj fxhp: salt-cloud 0.8.11
16:01 alunduil joined #salt
16:01 pdayton joined #salt
16:03 generj forrest: that's exactly it, so the fix is…
16:03 generj "curl -L 'http://bootstrap.saltstack.org' | sh -s -- -P -U -M -N git develop" ?
16:03 generj I'll try
16:03 generj weird that 2014.01 is taking so long
16:03 forrest generj, might be better to just pull from develop. I don't know if those options will install salt too
16:04 forrest generj, joehh worked on it last week I believe.
16:04 forrest salt-cloud worked fine for me in 0.17.4, haven't tested it on 0.17.5 yet
16:05 timoguin salt-cloud 0.8.11 still works for me on 0.17.5
16:05 forrest timoguin, good to know
16:05 timoguin it doesn't have the MS Azure provider though
16:05 timoguin I think that's 2014.1.0+
16:06 forrest azure, where you go if you want to spend even more money than amazon, with even worse support from Microsoft! :P
16:07 timoguin I'm looking at MS's BizSpark program. Gives start ups 100% free access to Azure, and all of MS software for three years
16:07 timoguin they're doing some interesting stuff
16:07 forrest timoguin, that's just so they can get you then charge 40% more than amazon's already outrageous prices
16:07 generj forrest: I'm using azure for free, thanks to bizpark, I get $5,000/mo to burn
16:07 timoguin by that point my infrastructure will bbe mobile and i can move it. :-P
16:08 generj timoguin: so you're saying the azure provider is 2014.1.0+ only and not in the cards?
16:08 sroegner_ joined #salt
16:08 danielbachhuber joined #salt
16:08 timoguin generj, yes, it is not in the 0.8.11 version of salt-cloud, which is what works with 0.17.5
16:09 timoguin it's in 2014.1.0 though
16:10 generj timoguin: ok, well at least I was correct about needing to upgrade to 2014.1, but there are no install guides!
16:10 generj timoguin: I can't find where to get it, I don't need a package just tell me what to do :)
16:10 fxhp generj: that string you posted with curl, that didn't do it?
16:10 fxhp generj: personally I'm waiting for the PPA to release
16:11 babilen fxhp: Okay, thanks. That works nicely now. :)
16:12 timoguin interesting, looks like all the docs for for salt-bootstrap had the curl | sh bits removed.
16:12 kballou joined #salt
16:13 babilen Another thing: I am using the users formula and have to set a shell which isn't installed by default. What would be the best way to enforce installation of that shell? I have a state foo.users and think that a requisite on the corresponding shell state would be an idiomatic thing here. Does that sound sensible?
16:13 ndrei joined #salt
16:13 timoguin babilen, requiring the state that installs the shell in the users state would work.
16:15 fxhp babilen - create an sls to install that shell
16:16 generj joined #salt
16:16 fxhp whenever you create a user that needs the shell, make sure it requires the shell
16:16 fxhp pkg
16:20 johtso joined #salt
16:21 ipmb joined #salt
16:24 generj fxhp: I can't wait for the PPA, I need this now :)
16:24 frasergraham joined #salt
16:24 Katafalkas Any news about videos from saltconf2014 ?
16:24 generj fxhp: the curl string got me Salt 2014.1.0-2476
16:24 generj but running sudo salt-cloud now gives me:
16:24 generj salt-cloud: error: No module named libcloud.compute.types
16:24 smcquay joined #salt
16:25 timoguin generj, salt-cloud was merged into salt in 2014.1, so there are normal modules and states now
16:25 generj its probably missing apache-libcloud
16:25 timoguin actually, i think salt-cloud is still an alias or something
16:26 timoguin but you may want to try running it how this doc shows: http://docs.saltstack.com/ref/modules/all/salt.modules.cloud.html
16:26 timoguin i would think it would have installed libcloud automatically, since that's listed in the requirements.txt for 2014.1
16:26 generj timoguin: thanks for that link! I'll try, but I just checked, and I am missing apache-libcloud indeed
16:28 linjan_ joined #salt
16:28 zooz joined #salt
16:29 generj timoguin: I would think so to, but nope, so I grabbed pip, and installed apache-libcloud and now it works
16:29 generj lets see how it deals with azure
16:29 timoguin gl!
16:30 timoguin generj, lemme know how it goes. i want to play with azure too
16:32 yomilk joined #salt
16:35 superdug left #salt
16:36 JaredR joined #salt
16:36 kermit joined #salt
16:36 symroe We at the Ministry of Justice in the UK have just released our libcloud provider for VMWare vCloud: https://github.com/ministryofjustice/salt-cloud-provider-vcloud
16:36 bemehow joined #salt
16:36 symroe Feedback would be great :)  Not sure how much use it will be to others, but I hope it will be
16:38 timoguin symroe, very cool!
16:39 ndrei joined #salt
16:41 timoguin symroe, it'd probably be quite easy to get that integrating into salt if you send a pull request
16:41 grep_awesome joined #salt
16:43 pdayton joined #salt
16:43 forrest hmm, has anyone seen where the salt master says the public key for a minion is cached, but the system doesn't show up in the unaccepted keys?
16:44 res0nat0r joined #salt
16:46 hunter joined #salt
16:46 forrest ahh nevermind, the dude screwed up his host data.
16:48 hunter_ joined #salt
16:48 Gareth morning
16:48 forrest hey Gareth
16:49 forrest were you spamming twitter this weekend from an app you wrote?
16:49 symroe timoguin: Yeah, we're going to add to it over the next few weeks, it's still not really stable.  Then we'll pull request it
16:49 generj joined #salt
16:50 forrest symroe, very could.
16:50 forrest *cool
16:51 Gareth forrest: not on purpose. wasn't an app I wrote.  I was testing out an app called Buffer.  Only sent one tweet.  were there more?
16:51 forrest Gareth, nope just the one
16:54 generj timoguin: no luck!
16:54 timoguin generj, :(
16:54 generj timoguin: sudo salt-cloud --list-locations my-azure-config
16:54 generj [INFO    ] salt-cloud starting
16:54 generj [WARNING ] The cloud driver, 'azure', configured under the 'my-azure-config' cloud provider alias was not loaded since 'azure.get_configured_provider()' could not be found. Removing it from the available providers list
16:54 generj Error: There was an error listing locations: No cloud providers matched 'my-azure-config'. Available selections:
16:54 generj just like before!
16:54 generj timoguin: I'm following this guide: http://docs.saltstack.com/topics/cloud/azure.html
16:54 ipalreadytaken joined #salt
16:56 generj joined #salt
16:56 forrest symroe, are you guys gonna keep the MIT license, or do an apache license like the rest of salt?
16:57 vbabiy joined #salt
16:58 timoguin generj, do you have the Azure Python SDK installed?
16:58 rallytime joined #salt
16:59 che-arne joined #salt
16:59 symroe forrest:
16:59 saurabhs joined #salt
16:59 symroe forrest: Good question
16:59 symroe I'll find out
17:01 forrest symroe, ok cool, it's not a big deal since MIT is a good license, just the more we can keep them the same the better!
17:01 generj timoguin: FINALLY! pip install azure did it!!!
17:01 jeremyfelt joined #salt
17:01 jdenning joined #salt
17:01 generj the final missing piece
17:01 timoguin there we go. :)
17:02 symroe forrest: I expect that would be required for a pull request in to salt though, so we should look in to it
17:02 Gifflen joined #salt
17:04 dyim joined #salt
17:04 timoguin generj, FYI, many modules will check that their dependencies are installed and Salt won't load them if they aren't.
17:05 generj timoguin: that's great, its just that the docs are basically misleading in terms of versions/install instructions and the curl thing doesn't pull everything in
17:05 generj also I sadly wasted a lot of time trying to get salt 0.17/18/whatever and salt-cloud going because their docs list azure support
17:06 schimmy joined #salt
17:06 timoguin yea i gotcha.
17:07 timoguin if there are docs for 0.17 that say there is azure support we should open an issue to fix that
17:12 forrest timoguin, http://docs.saltstack.com/topics/cloud/azure.html looks like the docs don't have a versionadded.
17:12 rgbkrk joined #salt
17:12 tharkun joined #salt
17:15 KyleG joined #salt
17:15 KyleG joined #salt
17:16 rainmanjam joined #salt
17:16 elfixit joined #salt
17:22 JasonSwindle joined #salt
17:23 HeadAIX joined #salt
17:25 rgbkrk joined #salt
17:27 forrest Anyone remember off-hand what Salt bases the key creation of a minion on? I've got two machines with different hostnames which are saying they have the same key, but the only matching items are in a few minor files from what I coworker is saying
17:29 dyim_ joined #salt
17:30 dave_den forrest: http://docs.saltstack.com/topics/tutorials/walkthrough.html#minion-id-generation
17:30 forrest dave_den, yea the hostname value is different.
17:30 forrest as is the fqdn
17:30 forrest I wasn't sure if there was somewhere else
17:31 mgw joined #salt
17:32 forrest dave_den, thanks, I'll tell my coworker to keep investigating since they're the one that seems to have caused the conflict in the first place :P
17:33 dave_den you break it you buy it, i suppose :)
17:35 forrest dave_den, heh yep
17:35 forrest I think it has to do with the fact the template he is using is based off that machine. So maybe something was pushed prior to hitting the master.
17:36 felixhummel joined #salt
17:38 Networkn3rd joined #salt
17:46 farra joined #salt
17:46 rgarcia_ joined #salt
17:47 rgarcia_ joined #salt
17:52 ckao joined #salt
17:52 akoumjian joined #salt
17:52 schimmy joined #salt
17:54 Tekni joined #salt
17:55 Philip joined #salt
17:59 xmltok joined #salt
18:02 dyim joined #salt
18:03 hunter_ joined #salt
18:04 Gifflen joined #salt
18:05 Gifflen joined #salt
18:07 schimmy joined #salt
18:07 Rasathus joined #salt
18:07 rallytime joined #salt
18:07 hunter_ joined #salt
18:08 ml_1 joined #salt
18:10 Rasathus Hello. I was wondering if anyone had any examples of their implementations using salt.states.network .  From the docs It looks like i need to write a node specific sls for each of my machines, but im not quite sure how to go about mapping to it from my top.sls   I know I can use the grains['id'] to target it, but im not sure how to do it without a folder for each node
18:10 timoguin forrest, sent a PR to fix the versionadded directive for Azure. :)
18:11 Rasathus Ideally I'd like a network folder with a number of sls files in a structure like <host id>.sls
18:11 ndrei joined #salt
18:13 mgw joined #salt
18:13 forrest timoguin, sweet
18:13 yomilk joined #salt
18:14 tom[] left #salt
18:15 kermit joined #salt
18:16 amckinley joined #salt
18:17 zooz joined #salt
18:19 forrest How does ansible have so many more stars on github than we do
18:21 timoguin I don't really star things. I just fork them.
18:21 timoguin Fork errythang.
18:21 timoguin number of forks is closer, but Ansible still wins
18:22 ewong_ joined #salt
18:24 forrest timoguin, star damn you star!
18:24 UtahDave joined #salt
18:24 sroegner_ joined #salt
18:25 naiyte joined #salt
18:26 forrest sroegner_, can you please merge this: https://github.com/saltstack-formulas/template-formula/pull/3
18:27 timoguin forrest, I did. :)
18:27 forrest when?
18:27 forrest oh you're talking about the star
18:27 forrest gotcha
18:27 viod1 hi! when highstate.show_highstate outputs "- shell: usr['shell']", it means it's not evaluated, right?
18:28 forrest sroegner_, also https://github.com/saltstack-formulas/users-formula/pull/19
18:28 sroegner_ forrest: ok if i do it in an hour?
18:29 forrest sroegner_, yea whenever
18:29 sroegner_ :-)
18:29 forrest they've just been hanging out for a few days now and I'm tired of seeing them.
18:29 MedicalJaneParis joined #salt
18:30 dyim joined #salt
18:31 sroegner_ right you are
18:31 sroegner_ ... done
18:31 forrest thanks man
18:32 sroegner_ anytime
18:34 kainswor I'm having some issues with 2014.1 on ubuntu
18:34 xmltok joined #salt
18:35 kainswor I thought perhaps it was my packaging, but after using setup.py to install it from the 2014.1 tag in the repo I get a StopIteration exception
18:35 kainswor relevant detail: I have a multi-master setup. I'll try to configure as a single master to see if it will work otherwise
18:35 kainswor Is this a known issue?
18:37 kainswor (I've confirmed single master works)
18:38 forrest kainswor, looks like there was https://github.com/saltstack/salt/issues/10732
18:38 forrest cachedout already fixed it.
18:39 kainswor as sweet, thanks
18:39 forrest or at least put in a merge which seems to fix it, might be worth testing if you can in a dev environment.
18:46 bt joined #salt
18:46 Gifflen joined #salt
18:47 jams joined #salt
18:49 jams Having an issue with my pillar data getting into my template when using salt-ssh. I shouldn't need to do anything special when using salt-ssh right?
18:50 kainswor yeah, I almost put the try: except in before myself, but that seemed too easy
18:50 dyim jams: this has been driving me crazy too. I see the pillar items but it isn't populating my templates
18:50 kainswor I'm packaging it up now, I'll see how it works
18:51 jams dyim: any workarounds?
18:51 dyim jams: not yet. I'm still trying to get it to work. I'm running on 0.17.5, you?
18:52 jams 2014.1.0
18:53 jams dyim: did you end up just installing the salt-minion on the node or what?
18:53 dyim jams: i'm trying to do everything via salt-ssh, trying to keep the nodes stripped down
18:54 bhosmer joined #salt
18:54 jams dyim: same. Did you find a bug report for this yet? Or any other notes/irc from others?
18:54 mgw1 joined #salt
18:54 kainswor are versions 2014.1 and 17.5 compatible?
18:55 forrest kainswor, they should be.
18:55 dyim kainswor: Yesterday in the irc channel, there was an issue related to what you were going through and the answer is no.
18:55 dyim kainswor: i think the server was 17.5 and minion was 2014.1
18:56 forrest dyim, really? I didn't think there was an issue.
18:56 dyim kainswor: it might work the other way around where the server was newer: 2014.1 and the minion was older
18:56 dyim forrest: it was driving these guys insane last week on this channel.
18:57 forrest UtahDave, can you confirm? Is there an issue with a 2014.1 minion talking to a 0.17.5 master?
18:57 forrest dyim, hmm, I must have missed it, been busy lately
18:57 dyim one was on epel and the other was epel-testing repo
18:57 swa_work joined #salt
18:57 forrest if there is an issue though, we need to document it
18:58 kainswor I get the test functions (those almost never break from version changes) but getting pillar info just fails with no good reason, and the state system can't find any states
18:58 dyim the issue was so severe, it didn't read the top file
18:58 forrest weird
18:58 dyim yup very weird
19:00 dyim jams: i didn't file a bug for this issue yet
19:00 dyim jams: still trying to understand the issue. I'm new to salt
19:00 diegows joined #salt
19:02 jams dyim: ya this bug also makes the salt-ssh feature pretty unusable for me too
19:03 Gareth Anyone heard any updates on the status of the debian/ubuntu packages for 2014.1.0?  See debs are in the repo but not referenced in the available packages.  Also looks like a dependancy isn't met on Debian stable.
19:03 xDamox In SaltStack, is there a directive similar to require to ensure a state has already been executed and it was successful?
19:13 viod1 I'm still struggling with my issue fromo several hours ago, could someone help me find the syntax for accessing attributes from a pillar in a for loop? The relevant files are pasted here: http://pastebin.aquilenet.fr/?24df2bf8c514d913#JlECrysGlirU32dtkdZWfQAiExc46+4fAzud7MShIGs= , and my issue is described more precisely at the bottom of the page. Thanks in advance!
19:13 n8n joined #salt
19:17 Gareth viod1: this is what I'm using to pull values from pillar and use them in a for loop. salt['pillar.get']('bind:configured_zones', {}).iteritems()
19:19 viod1 Gareth: thanks, but i still get the same error :(
19:21 Gareth usr['home'] didn't work?
19:22 viod1 nop
19:22 Gareth what error did it give you?
19:23 viod1 "Jinja variable 'list object' has no attribute 'home'" (it's all written in the pastebin)
19:28 Gareth viod1: looks like it *should* work.  You're using iteritems and usr['home']
19:28 Gareth ?
19:28 kermit joined #salt
19:29 viod1 yep
19:29 viod1 {% for name, usr in salt['pillar.get']('config_users', {}).iteritems() %}
19:29 viod1 - home: {{ usr['home'] }}
19:30 viod1 my code looks a lot like what i've found on the internet, i really don't understand why it doesn't behave as expected
19:30 amckinley joined #salt
19:32 Gareth viod1: looks similar to what I'm using for a state.  which works.
19:33 viod1 could it be some kind of version problem?
19:33 viod1 I'm using the debian stable repo, and nothing installed except for salt (and the dependencies automatically installed by apt-get)
19:34 viod1 the salt repo for debian stable, I should say
19:35 dyim jams: I found this: https://github.com/saltstack/salt/issues/8340
19:35 dave_den viod1: your user attributes are formed as a list.
19:37 vu joined #salt
19:37 jams dyim: good find.
19:37 dyim jams: https://github.com/saltstack/salt/issues/8733 this but seems to be present in 0.17.5, they said it should be in 2014.1.0 but not sure
19:37 dyim jams: according to you, it doesn't work in 2014.1
19:38 viod1 dave_den: indeed, it works without the dashes. I never knew when I had to put them or not, now I will ^^
19:38 viod1 Gareth: dave_den: thanks a lot for your help :D
19:38 jams dyim: right, still an issue in 2014.1.0
19:39 ml_1 joined #salt
19:39 Gareth dave_den: hah. I completely missed that part.  thanks.
19:43 joshk joined #salt
19:43 timoguin_ joined #salt
19:44 erchn joined #salt
19:44 erchn hi there...anyone using External Auth out there with PAM?
19:45 martoss joined #salt
19:45 erchn I've got it working well, until I try and use "--subset=1" or similar subset arguments, it throws an auth error
19:46 erchn I couldn't find an issue filed for it, running 0.17.5
19:46 farra joined #salt
19:54 erchn hmm, trace level debug offered no clues :-/
19:57 musonic joined #salt
19:57 musonic Hi, I'm trying to configure my vagrant vm using salt but I seem to be having problems getting php to install correctly
19:58 musonic if i type php -i i get and error saying that php is not currently installed.
19:58 ipalreadytaken joined #salt
19:59 musonic is that just because the cli isn't installed?
20:00 dave_den musonic: in debian based distros like Ubuntu you will need to install php-cli to have the 'php' command available
20:00 vejdmn joined #salt
20:01 Philip_ joined #salt
20:01 musonic dave_den: that's what i thought, so I added it to the list of the packages requested in my php55.sls file
20:01 mgw joined #salt
20:03 Philip_ Hey guys, just recently upgraded from .17 to latest and I'm seeing all my pkg.installed stats fail that have 'enablerepo: epel', comments coming back along these lines: 'Comment: Package 'npm' not found (possible matches: npm)'
20:03 dyim jams: i'm going to try 2014.1.0rc3
20:03 Philip_ trying to track down where the available packages failure is happening, but I thought perhaps there would be something major that changed that I'm unaware of
20:04 dyim that is where thatch commited the fix
20:05 jams__ joined #salt
20:07 shadowsun Hey
20:07 shadowsun Did 2013.1.0 get pushed to release inspite of the multimaster bugs?
20:07 shadowsun 2014.1.0 even
20:07 bhosmer joined #salt
20:08 ajw0100 joined #salt
20:10 forrest shadowsun, it was already packaged before that bug was discovered.
20:10 shadowsun basepi: Yes, I can test the patch, because omg my production environment.. sec
20:11 shadowsun forrest: yeah, but it was in testing, I didn't think it would go through to epel normal
20:11 forrest shadowsun, well, the bug was only filed 3 days ago
20:11 forrest and it's been in epel for a few weeks
20:11 hunter_ joined #salt
20:11 forrest or a week and a half
20:11 forrest because, lol epel
20:11 basepi shadowsun: yes, we will be releasing a 2014.1.1 in the near future with the multimaster fix.
20:12 shadowsun yeah
20:12 shadowsun I'm rolling zabbix into the environment to alert me if this happens again
20:12 shadowsun because I should have caught it earlier :|
20:13 ajw0100 joined #salt
20:15 akoumjian joined #salt
20:15 seblu joined #salt
20:15 Gordonz joined #salt
20:16 Gordonz joined #salt
20:16 yomilk joined #salt
20:19 joshk I'm using the salt python module to ssh to a bunch of hosts to delete users. The auth setup on these machines is not 100% correct, so on some machines it immediately fails and prints 'permission denied (publickey)'. Then, the error from salt is really obtuse (KeyError: 'grains' as it tries to parse the output) and salt eventually hangs
20:20 joshk is that a known issue?
20:20 shadowsun basepi: This fixes the obvious problem; but now the minion doesn't ever come fully online
20:20 shadowsun [DEBUG   ] I am art.artstreetaustin.com and I am not supposed to start any proxies.
20:20 forrest shadowsun, that domain...
20:21 shadowsun >.> was supposed to be munged
20:21 basepi shadowsun: o.O
20:21 forrest well, it's written forever now
20:21 forrest shadowsun, now you can change it to not sound weird, and have it be streetart
20:22 shadowsun lol
20:22 shadowsun it's not a domain I picked if that helps
20:23 forrest a little
20:23 elfixit joined #salt
20:23 basepi shadowsun: that log doesn't point to any problems.  it logs that if your minion is not configured as a proxy minion.
20:23 shadowsun basepi: okay, that's where it stops
20:23 shadowsun sec
20:23 basepi so it will log it in the general case.
20:24 shadowsun I'm going to run them both in debug and see if I get anything
20:26 shadowsun basepi: I'm not getting any logs of failures, but the minion with this patch just doesn't do anything.
20:26 shadowsun Minion did not return
20:26 shadowsun salt-run manage.down still shows it as offline even.
20:26 basepi shadowsun: very strange.  can you do like a `salt-call test.ping` on the minion?
20:26 shadowsun Whatever it's doing, it's not communicating with the master
20:26 shadowsun yes, sec
20:27 Heartbane so I just deployed a new minion running the 2014xxxx version of salt but the master is still running 0.17.5 tried to push a state and I get "Data failed to compile:" yet it works on all minions running 0.17.5?
20:28 basepi Heartbane: master needs to have a version >= minions
20:28 shadowsun https://gist.github.com/Kionmaru/49d9cc389a42ff10f8ee
20:28 Heartbane basepi: thanks
20:28 shadowsun basepi: it can ping itself, it looks like.
20:28 forrest basepi, can you add that to the 2014.1.0 release notes. I wasn't aware there was going to be that sort of compatibility issue again like with 0.16 -> 0.17
20:28 forrest wasn't documented anywhere.
20:28 basepi forrest: i actually didn't either
20:29 basepi though in general we don't guarantee compatibility with master < minion
20:29 * Gareth shoots basepi a disapproving glance
20:29 basepi it's much more important to document hard incompatibilities
20:29 * Heartbane nods at Gareth.
20:29 basepi Gareth: ?  just upgrade your master first is the general best-practice
20:30 basepi of course, every once in awhile a security issue will cause an incompatibility like 0.17.0 and 0.17.1 did.
20:30 Heartbane basepi: no worries I will just blame UtahDave
20:30 shadowsun lol
20:30 forrest basepi, yea the security issue was understandable, but this time there isn't any sort of notice like that.
20:30 UtahDave :(
20:31 Gareth poor UtahDave
20:32 * Heartbane UtahDave hey I have my salt-minions deploying openldap for credentials on all my guests. Pretty slick
20:32 UtahDave woot!
20:32 Heartbane That was supposed to be /msg not /me ..... I hate Monday nothing werqz right
20:32 forrest no pity for UtahDave, he gets to fly all over the world
20:33 basepi forrest: i agree.  it's less than ideal
20:33 Gareth basepi: Oh I'm not disapprovaling of that...but I'm sure you've done something that requires a disapproving glance.
20:33 basepi in fact, i have no idea why they're not compatible
20:33 basepi theoretically it should be fine.  I want to look into that When I Find Some Time (tm)
20:33 steven_schwartz joined #salt
20:33 forrest basepi, yea I didn't know until this morning when people started complaining, and apparently it was an issue for someone last week
20:33 forrest basepi, heh, use your magic wand, freeze time and space, done
20:34 shadowsun basepi: You get that gist/
20:34 steven_schwartz OK. Weird circumstance -- I went from having a working system, to one that isn't loading anything. I get, when I try to run a state.highstate: No Top file or external nodes data matches found
20:34 basepi shadowsun: i did.
20:34 Philip_ @steven_schwartz what does your /etc/salt/master look like?
20:34 basepi shadowsun: so next thing i'd like you to do is start the minion while you have the master log up in `tail` or something similar to see if the master has any log activity when the minion starts.
20:35 basepi shadowsun: trying to figure out what the communication problem is.
20:35 steven_schwartz Which parts do you want to see, Philip_?
20:35 jeremyBass1 hello, is there any reason with the new updates out that the pillar data would be proccessed out of order from the order it's set in the pillar file ?
20:35 steven_schwartz file_roots:   base:     - /srv/salt
20:35 Philip_ your file_roots and pillar_roots
20:35 shadowsun basepi: it does, I'm in debug already
20:36 steven_schwartz I don't have pillar_roots changed from the default, and there's nothing in the pillar. But it was working before without touching pillar.
20:36 steven_schwartz And there is a top.sls in /srv/salt.
20:36 Philip_ gotcha, and there is a file at /srv/salt/top.sls?
20:36 Philip_ k
20:36 Philip_ what happens when you restart the master and tail the master log?
20:37 themadcanudist joined #salt
20:37 shadowsun basepi: reload the gist
20:37 bluetorch joined #salt
20:37 steven_schwartz When not in debug mode, nothing. Just a notice of the stop -- oh, wait! 2014-03-03 20:37:47,418 [salt.master                                 ][WARNING ] Authentication failure of type "user" occurred.
20:38 themadcanudist hey guys… is there any way to install the mako packages on destination hosts via a highstate WITHIN the same run that needs mako? (all my formulas use mako)
20:38 basepi shadowsun: so it seems to have auth'd just fine.  what does `salt 'exa.example.com' -v test.ping` return?
20:38 themadcanudist mako pip package*
20:38 martoss joined #salt
20:39 steven_schwartz No minions matched the target. No command was sent, no jid was assigned.
20:39 shadowsun basepi:     Minion did not return
20:39 steven_schwartz Whoops -- ansering the wrong question. ;)
20:39 basepi shadowsun: and was there any log activity on the minion?
20:39 shadowsun basepi: none.
20:40 analogbyte is there a way to use a returner by default for every job?
20:40 steven_schwartz The only recent change I'm aware of was setting the number of worker threads to 10, then resetting it to 5.  I don't think this would have caused such an issue, but...
20:40 basepi shadowsun: hrm....
20:40 Philip_ Hmm I've never seen that error before, perhaps someone else on here will have seen it
20:41 basepi shadowsun: if you change it to single master instead of multi-master, does it work?
20:41 basepi shadowsun: want to make sure it's related to multi-master before we go further.
20:43 Philip_ gotta run, sorry steven
20:44 shadowsun basepi: lets double check it, goo idea
20:45 jeremyBass1 any one have an idea why pillar data order is not unsorted?  I am using the bootstrap, and noticed to day i needed to change {{ env }} but now the pillar data is out of order too..
20:45 shadowsun basepi: it's highstating, which is a lot more than it was doing
20:45 shadowsun basepi: tons of things working right now.,
20:45 basepi shadowsun: you have it running startup_states?
20:45 shadowsun yes
20:46 shadowsun and I can now test.ping it from the salt-master.
20:46 shadowsun So yeah, this appears to still be related to multi-master config
20:47 basepi shadowsun: two things to verify:  multi-master was working for you in 0.17.5, right?
20:47 shadowsun basepi: Yes, our entire config has it
20:47 shadowsun 40+ minions
20:47 basepi shadowsun: kk.  can you try re-enabling multi-master but disabling startup_states?
20:47 * hunter has reclass working
20:47 shadowsun I can. one second.
20:47 basepi shadowsun: ty
20:48 shadowsun basepi: yw.
20:48 Pate_ joined #salt
20:49 shadowsun basepi: multi-master without startup_states exhibits the same non-returning behavior.
20:49 ipalreadytaken joined #salt
20:49 basepi shadowsun: thanks.  just wanted to make sure.
20:49 shadowsun basepi: I like making sure.
20:49 basepi shadowsun: I'm going to get Mike (cachedout) to spin up his environment again and do some multi-master testing, see if we can track down what the problem is.
20:50 shadowsun basepi: If I can help let me know.
20:50 tmihai joined #salt
20:50 shadowsun I'll keep an eye on this
20:50 basepi shadowsun: will do.  thanks for the patience.
20:50 shadowsun basepi: Our environment is mostly dev, so it's designed to catch this stuff
20:50 shadowsun and the majority of our production environment is between uses
20:51 shadowsun so >.> I have a work around I can use for now, but I hate not having redundancy
20:52 gnugnu8 joined #salt
20:52 mgw1 joined #salt
20:52 steven_schwartz Anyone else here seen an error like: No Top file or external nodes data matches found ?
20:53 steven_schwartz Because it has managed to completely paralyze my entire salt infrastructure.
20:55 robawt joined #salt
20:56 alexthegraham joined #salt
20:56 jeremyBass1 is there a new way to sort pillar data now?  I have been counting on the order of the pillar data to controll the order of installing plugins and now today it's all out of order when i run salt, and even see that it's store the way it was run when i do a salt-call pillar.items
20:56 Guest72558 Hmmm
20:56 alexthegraham Quick (hopefully) question on states.pkg: How are version numbers assessed?
20:56 Guest72558 How do I debug the following error message?
20:56 Guest72558 No matching sls found for 'iptables' in env 'base'
20:57 mgw joined #salt
20:57 steven_schwartz What does your file_root look like?
20:57 shadowsun basepi: it's not multi-master.
20:57 alexthegraham If I tell pkg.installed to look for 2014.1.0, and I have 2014.1.0-1.el6 installed, it tells me it needs to be installed. But grains.get saltversion has 2014.1.0.
20:57 steven_schwartz Guest, do you have either an iptables.sls or an iptables directory under your file_root with an init.sls?
20:57 shadowsun basepi: check this out: doing ^master:\n - master\n with a single minion gives the same behavior.
20:58 ipalreadytaken joined #salt
20:58 shadowsun basepi: Is there any way to configure multiple masters on a single line?
20:58 shadowsun s/single minion/single master/
20:59 scarcry joined #salt
21:00 kintel joined #salt
21:01 basepi shadowsun: so the difference is that the master is formed as a list instead of just a string, and that triggers the multi-master behavior.
21:01 basepi shadowsun: and yes, we're definitely going to fix this, multi-master support is important for redundancy.
21:02 shadowsun yeah
21:02 shadowsun I know we're going to, if I can help I want to
21:02 basepi shadowsun: but ya, doing multiple masters on a single line wouldn't help, because you still have to form them as a list which will trigger the behavior (and the accompanying bugs)
21:02 shadowsun If I had time I'd be in the source code trying to figure it out myself but I'm in the middle of teaching a class
21:02 shadowsun I was afraid of that :(
21:03 basepi shadowsun: cachedout is knee-deep in a syndic bug at the moment, hopefully he'll finish it soon and can revisit the multi-master issue.
21:03 shadowsun basepi: Cool. I'm going to make some changes to my environment, enough to work around this for now
21:04 basepi shadowsun: cool.
21:04 basepi shadowsun: glad you have a workaround
21:04 valgrind joined #salt
21:07 shadowsun basepi: me too!
21:09 mgw joined #salt
21:10 steven_schwartz Sorry to be a pest, but right now I'm completely hosed: I have a top.sls in my file-root, and keep getting: No Top file or external nodes data matches found -- has anyone else seen this? I am asking again because right now, salt is completely hosed as a result. I can salt 'base-app-1' test.ping, and get a result, but not a state.highstate.
21:12 steven_schwartz (Similarly, trying to invoke any other states in the file_root result in "No matching sls found for 'nagios' in env 'base'" or the like, even though /srv/salt/nagios/init.sls exists.
21:14 erchn steven_schwartz: is there any templating in your top file that could be borked?
21:15 steven_schwartz Would a syntax error in any subtemplate bork the whole thing?
21:15 steven_schwartz Right now, all it has is a set of grain matches.
21:15 erchn you can run "state.show_top" without error?
21:16 steven_schwartz Nompe. Python errors a-go-go.
21:16 steven_schwartz What target should I be using? Not just salt state.show_top, or?
21:16 erchn I missed what version you were running
21:17 erchn try: salt 'base-app-1' state.show_top
21:17 steven_schwartz base-app-1:     ----------
21:17 erchn that's it?
21:17 steven_schwartz Ayup.
21:17 erchn sounds like a yaml syntax issue in the top file
21:18 erchn maybe try pasting it into a yaml lint tool
21:18 raizyr joined #salt
21:19 helderco joined #salt
21:19 pdayton joined #salt
21:19 steven_schwartz salt 0.17.2, FYI.
21:19 erchn yamllint.com shows down for me, you could try http://yaml-online-parser.appspot.com/
21:20 steven_schwartz Tried that, it shows no problem.
21:20 steven_schwartz At least, I presume it would mark an error.
21:20 erchn it should, yea
21:20 steven_schwartz Yeah, no error.
21:21 erchn are all hosts broken or just one?
21:21 steven_schwartz All hosts of 1 class is broken. I'm going to try and build one of the other.
21:22 erchn you could try clearing out the state cache on one of those hosts
21:22 erchn maybe a caching issue
21:24 steven_schwartz salt 'base-app-3' state.clear_cache produces just base-app-3:. I tried running a subcommand: salt 'base-app-3' state.sls nagios -- and I get: base-app-3:     Data failed to compile: ----------     No matching sls found for 'nagios' in env 'base'
21:24 steven_schwartz As far as I can tell, there's a problem finding the base directory.
21:25 ipalreadytaken joined #salt
21:25 rgbkrk joined #salt
21:26 schimmy joined #salt
21:27 steven_schwartz And, indeed, verified: I tried with my other profile, got the same result.
21:27 ipalread_ joined #salt
21:29 erchn and you're sure the file_root section for base isn't wrong?  server's disk full? :-)
21:30 schimmy1 joined #salt
21:30 steven_schwartz Filesystem            Size  Used Avail Use% Mounted on /dev/xvda1            7.9G  4.5G  3.4G  57% /  ;-) Believe me, I wish it were disk full.
21:30 steven_schwartz file_roots:   base:     - /srv/salt/   dev:     - /srv/salt/dev/services     - /srv/salt/dev/states   prod:     - /srv/salt/prod/services     - /srv/salt/prod/states
21:30 steven_schwartz So, the top file should be, in this model, /srv/salt/top.sls, right?
21:31 steven_schwartz -rw-r--r-- 1 root root 307 Mar  3 21:06 /srv/salt/top.sls
21:32 jeremyBass1 gosh, does anyone at all know why the pillar data for my set up is not being loaded out of order from the order i set in the pillar file?  for a long while it was saying in the order i put in the pillar file, now, it's not keeping that order anymore..
21:33 jeremyBass1 is being loaded out of order i mean*  it's not the same as the pillar file when i do a salt-call pillar.items like it use to be..
21:33 jeremyBass1 so when i loop of the data things are going out of order now
21:35 hunter jeremyBass1: Unless I'm mistaken (and I often am) the pillar data is just a dictionary in python, and python dicts have no order guarantee.
21:36 hunter jeremyBass1: the fact that you were getting a desirable order before was possibly an accident.
21:36 ipmb joined #salt
21:36 jeremyBass1 @hunter, well it's odd that it has been stable with edits and changes to the pillar file for months on end..
21:37 yusuket joined #salt
21:37 jeremyBass1 it's not even in a-z order or any order i can tell
21:37 jeremyBass1 oh wait i see it's not A-Z then a-z ..
21:37 jeremyBass1 well that is fun
21:38 hunter jeremyBass1: its in dict.c bucket order. :)
21:39 jeremyBass1 well so how to i enforce the order now that it's deciding to be in dict.c bucket order
21:39 jeremyBass1 it's very odd that it was saying in order for the whole time, till just a few days ago.. something had to have changed.. :-\
21:40 themadcanudist hey guys, is there a way to install mako during a high state, if that high state includes templates that need to be parsed by mako?
21:40 erchn steven_schwartz: I'm not sure, very odd.  can you start the salt-master with -l debug and see if anything on the console shows obvious issues?
21:41 Ryan_Lane joined #salt
21:42 jeremyBass1 brb but if anyone has an idea on how to fix this issue I'm wide open
21:44 steven_schwartz I'm trying, and watching it scroll by.
21:45 steven_schwartz Well, max_open_files was set too high, set down to default. I doubt that's the issue.
21:46 steven_schwartz Missing configuration file: /root/.salt
21:46 elfixit joined #salt
21:46 martoss joined #salt
21:46 steven_schwartz Skipping various things as it starts up, nothing immediately obviously related.
21:47 cewood joined #salt
21:48 steven_schwartz Accepting a bunch of authentication keys...
21:48 steven_schwartz Now getting paylod commands for payloads that didn't get delivered because of this problem...
21:49 steven_schwartz No noticeable errors even when I do a salt-call from a minion, trying to install, and getting a "no matching sls" error.
21:49 jeremyBass1 well the only thin i can find on the pillar data sort oder is https://groups.google.com/forum/#!msg/salt-users/fKpxJ9Q44oQ/2zr52AalXHEJ
21:50 kermit joined #salt
21:58 Pate joined #salt
21:58 vejdmn joined #salt
21:59 yusuket exit
21:59 yusuket left #salt
21:59 Guest29034 joined #salt
22:01 erchn steven_schwartz: nothing in dmesg like selinux errors or something?
22:01 steven_schwartz Nompe.
22:02 erchn what does "salt 'base-app-1' cp.list_master" show?
22:02 yomilk joined #salt
22:02 steven_schwartz base-app-1:
22:02 erchn looks like the salt master just cannot read your file_root
22:03 steven_schwartz Which is bizarre, given that it's running as root and the directory is 755.
22:03 steven_schwartz By default, it goes to /srv/salt, right?
22:04 erchn yea
22:04 erchn your /etc/salt/master or /etc/salt/master.d/*.conf file doesn't change that definition at all?
22:05 steven_schwartz I had uncommented the file_roots directive, but not altered it.
22:06 steven_schwartz Even with everything set to default, it's refusing to read /srv/salt/top.sls -- or anything in the "base" environment.
22:07 erchn can you strace the running salt-master and do a list_master, or state.show_top and see what file path it's trying to read?
22:08 steven_schwartz Well, this bloody system only has autrace, but I'll give it a shot.
22:09 steven_schwartz And it doesn't like it when I do that. (Salt-master appears to hang)
22:11 lzhang joined #salt
22:12 erchn beyond that, I'm not sure
22:12 erchn good luck...
22:12 steven_schwartz THank you. I've reported it as an issue, so we'll see.
22:14 seapasulli joined #salt
22:15 pdayton joined #salt
22:16 n8n joined #salt
22:17 cetex joined #salt
22:20 naiyte joined #salt
22:21 andrej joined #salt
22:23 andrej steven_schwartz : sorry, back w/ my own account now (must have been a netsplit I didn't notice) and yes, I have an iptables direcotry w/ an init.sls in it
22:23 andrej Also apologies for late response - I was dragged into a meeting
22:23 steven_schwartz Then I fear I cannot help you -- you've hit my limit of guesswork. (No problem.)
22:23 andrej Heh
22:24 andrej I have a lot of other, similar structures, e.g. ntp, nrpe, yum ... and all work
22:24 andrej just this iptables one doesn't want to
22:24 steven_schwartz I'd verify the init.sls yaml,
22:24 timoguin joined #salt
22:24 Gareth andrej: what fails with the iptables one?
22:25 andrej Gareth : I get Data failed to compile: No matching sls found for 'iptables' in env 'base'
22:26 steven_schwartz Heh. That appears to be catching today. At least it only fails for one for you.
22:26 andrej Oh
22:27 andrej Well .... I assumed it was still working; I just tried another one (didn't need to do ntp in 2 weeks), and it's now also failing :(
22:27 Gareth andrej: I saw that same error last night.  Your minion and master are the same version?
22:28 steven_schwartz Weird -- is it happening to just one module, or several, Gareth? (And what version are you running?)
22:28 andrej Hmmm
22:28 andrej the minions have an odd version number
22:29 andrej 2014.1.0
22:29 andrej master is 0.17.5
22:29 Gareth steven_schwartz: I was only testing one module.
22:29 forrest yea that doesn't work andrej
22:29 Gareth andrej: that is your problem :)
22:29 forrest andrej, master version has to be greater than minion
22:29 forrest we're just learning this today
22:29 Gareth what forrest said.
22:29 andrej I just ran a yum update this morning
22:29 forrest Gareth, was that documented anywhere that you know of?
22:29 forrest Colton wasn't aware it was a problem either, and frankly I don't see why it SHOULD be
22:29 Gareth forrest: I think it might have been mentioned somewhere.  don't remember what.
22:29 forrest there were no transport layer changes.
22:30 erchn I've seen that somewhere, cannot remember where...
22:30 Gareth forrest: I think it was one of those things that wasn't supported and frowned upon.
22:30 erchn maybe from Joseph in a training session
22:30 forrest Gareth, well, that was the case for 0.16 -> 0.17 because of the zeromq changes.
22:30 forrest there was no warning that the same would happen for 0.17.5 and 2014.1
22:30 Gareth forrest: perhaps thats where I saw it :)
22:30 forrest Gareth, You're supposed to have all the answers!
22:31 steven_schwartz That explains it -- I bet that is it, let me try updating my local master, because the clients are newer.
22:31 Gareth forrest: hah
22:31 erchn steven_schwartz, ahh, you told me 0.17.2!!
22:31 andrej Damn
22:31 erchn :-)
22:31 andrej there's no new version for ubuntu yet
22:31 cedwards joined #salt
22:31 steven_schwartz for the *master*, not for the *client*. I didn't think to check the *client*.
22:31 forrest andrej, I believe that joehh is still working on it?
22:31 andrej so how do I downgrade my centos minions? :)
22:32 erchn try running a "salt-call state.show_top" on the master
22:32 forrest andrej, try yum downgrade salt-minion
22:32 yomilk joined #salt
22:32 steven_schwartz It is worth noting that the test ping works between versions.
22:32 forrest andrej, http://docs.fedoraproject.org/en-US/Fedora/14/html/Software_Management_Guide/ch05s16.html there's also some docs that might help
22:32 harobed_ joined #salt
22:32 steven_schwartz Well, that's odd.
22:32 mgw1 joined #salt
22:33 fllr joined #salt
22:33 dangra_ joined #salt
22:33 fllr Hey guys. How can I use a salt pillar on a module call?
22:33 harobed_ joined #salt
22:35 steven_schwartz local:     ----------     base:         - dev         - appserver
22:35 steven_schwartz So that works.
22:36 steven_schwartz Is an upgrade just a yum upgrade salt-master?
22:36 steven_schwartz (And will that break my salt-cloud installation?)
22:36 andrej meh - no worky
22:38 faldridge joined #salt
22:39 hunter Our mirror here updated - I just had to make a copy of all the 0.17.5 packages....
22:39 hunter <--- not ready for 2014.1
22:39 zooz joined #salt
22:39 JaredRo I'd like to do a salt.highstate, and later, be able to go to a given minion and be able to figure out when the highstate happened, and where the source code came from for the sls files (git repo, branch, commit). Ideas?
22:41 steven_schwartz OK. When I try and do the upgrade, I get a conflict with /usr/bin/salt-cloud -- does anyone know if I need to do a remove and reinstall?
22:41 steven_schwartz I am nervous about this, but rather have to do it, as I do an automated build that pulls from an already-upgraded repo.
22:42 ndrei joined #salt
22:44 forrest steven_schwartz, salt-cloud is integrated with salt 2014.1, so yea you'd have to remove it. I'd really suggest to do it in dev if you can
22:44 steven_schwartz Fortunately, it's all dev. But the salt-cloud commands will continue to work, yes?
22:44 forrest once you reinstall 2014, they should
22:44 forrest I haven't messed with it yet
22:45 NotreDev joined #salt
22:47 timoguin joined #salt
22:47 steven_schwartz Well, wish me luck. ;)
22:47 forrest Sorry I need all this luck
22:47 pdayton joined #salt
22:47 forrest none for you
22:47 steven_schwartz Heh. :P
22:49 steven_schwartz Are we expecting to have this issue every time salt is upgraded, or just major versions? This will be a pain if I have to upgrade the master every single time in short order after the new code is pushed out. (I'm using this for cloud building, so the minions build with the newest EPEL repo.
22:49 manfred does anyone have a good way to refresh the salt python environment after pip installing something?  I am installing a library that is needed for a different module, but it isn't discovering the module until i manually go in and do a service salt-minion restart, saltutil.sync_all didn't help
22:50 manfred ahh reload_modules apparently
22:59 jeffrubic joined #salt
23:00 swa_work joined #salt
23:02 naiyte left #salt
23:02 naiyte joined #salt
23:03 ndrei joined #salt
23:08 mgw joined #salt
23:09 yomilk joined #salt
23:15 themadcanudist joined #salt
23:20 Ryan_Lane joined #salt
23:21 steven_schwartz And that solved the problem -- thank you, all.
23:22 yomilk joined #salt
23:24 srage joined #salt
23:24 shadowsun ...okay what
23:25 swa_work joined #salt
23:25 srage joined #salt
23:26 quickdry21 joined #salt
23:26 elfixit joined #salt
23:27 ipalreadytaken joined #salt
23:31 hardwire joined #salt
23:32 ndrei joined #salt
23:34 zdong does anyone have luck to put custom execution module on git ?
23:36 zdong my custom module works on local /srv/salt/_modules
23:36 zdong but on remote git .
23:37 themadcanudist hey guys, is there a way to have pip install mako on a minion when a high state is run before the parsing of any mako templates are performed? Sort of a bootstrapping?
23:38 joehh andrej: apologies for ongoing delay - expecting to finish all backports today (debian and ubuntu)
23:38 bemehow joined #salt
23:41 dave_den themadcanudist: i believe you can do it with a pydsl template as long as that tempalte state file is ordered before any mako templates
23:41 patrek_ what would be the best way to update a package through salt? (using centos if it makes a difference)?
23:41 dave_den themadcanudist: the sure-fire way would be to run a single state.sls to bootstrap the install of mako before running the highstate
23:42 hardwire cannot seem to find a good example of calling events.fire_master from an sls
23:42 hardwire or reactor
23:42 steven_schwartz You could run a run.cmd on all clients for the install command, patrek.
23:43 themadcanudist dave_den: why would it have to be a pydsl template that's ordered vs. any template that is ordered?
23:43 dave_den because pydsl templates have render time executiong capability
23:44 themadcanudist sorry, do you mind elaborating? What does that mean?
23:44 themadcanudist in the grand scheme of things?
23:44 dave_den themadcanudist: http://docs.saltstack.com/ref/renderers/all/salt.renderers.pydsl.html#render-time-state-execution
23:44 themadcanudist thx
23:45 themadcanudist aha
23:45 alunduil joined #salt
23:46 themadcanudist thank you sir! Much appreciated
23:46 dave_den good luck
23:47 Schmidt joined #salt
23:47 diegows joined #salt
23:53 conan_the_destro joined #salt
23:53 Nazca joined #salt
23:55 l0x3py joined #salt
23:58 UtahDave joined #salt

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