Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2016-08-25

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

All times shown according to UTC.

Time Nick Message
00:23 Egyptian[Home] joined #fuel
00:42 Julien-zte joined #fuel
01:04 Julien-zte joined #fuel
01:13 raunak joined #fuel
01:34 fatdragon joined #fuel
01:36 code-R joined #fuel
01:48 ilbot3 joined #fuel
01:48 Topic for #fuel is now Fuel 9.0 (Mitaka) https://www.fuel-infra.org/ | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
01:52 code-R_ joined #fuel
02:20 raunak joined #fuel
02:21 justif joined #fuel
02:28 thiagolib joined #fuel
02:30 code-R joined #fuel
02:30 serverascode joined #fuel
03:09 raunak joined #fuel
03:41 Julien-zte joined #fuel
03:46 raunak joined #fuel
03:58 raunak joined #fuel
04:19 makita_tats joined #fuel
04:29 raunak joined #fuel
05:00 raunak joined #fuel
05:23 pradiprwt joined #fuel
05:24 pradiprwt Hi Everyone, I am trying to copy some python files to fuel nodes using plugin but it is not copying. Can any refer me what exectly aproach I need to follow.
06:06 fatdragon joined #fuel
06:48 vkulanov joined #fuel
07:04 fatdragon joined #fuel
07:12 pradiprwt what puppet module source I need to use for copying files
07:13 pradiprwt I am using "puppet:///modules/demo/test.py"
07:36 pradiprwt fuel plugins module source is not in default like "/etc/puppet/modules", can any one please tell me what is module source for plugins ..????
08:05 fatdragon joined #fuel
08:10 yassine joined #fuel
08:19 skath joined #fuel
08:29 noshankus joined #fuel
08:43 aarefiev_ joined #fuel
08:44 azvyagintsev joined #fuel
09:20 skolekonov pradiprwt, hi, AFAIR you can find them in /etc/fuel/plugins/
09:21 pradiprwt skolekonov : if we want to copy any file from plugin module what module we need to define ..
09:22 pradiprwt is it "puppet:///modules/module_name/file_name"
09:23 pradiprwt I mean source in puppet
09:32 mskalski pradiprwt: here is an example https://github.com/openstack/fuel-plugin-contrail/blob/6684fc6c2c12ba3e43c46be3418330de1a95f786/deployment_scripts/puppet/modules/contrail/manifests/database.pp#L97  and file is located here https://github.com/openstack/fuel-plugin-contrail/tree/6684fc6c2c12ba3e43c46be3418330de1a95f786/deployment_scripts/puppet/modules/contrail/files
09:36 mskalski pradiprwt: and for puppet task you should define puppet modules like here https://github.com/openstack/fuel-plugin-contrail/blob/master/deployment_tasks.yaml#L255
09:44 pradiprwt Thanks mskalski , I will try this now
09:49 krypto joined #fuel
10:02 xarses_ joined #fuel
10:05 fatdragon joined #fuel
10:13 ddmitriev joined #fuel
10:47 Egyptian[Home] joined #fuel
10:48 cartik joined #fuel
11:08 Egyptian[Home] joined #fuel
11:09 fuel-slackbot joined #fuel
12:03 ilozgachev joined #fuel
12:04 ilozgachev Hey! How can I get Horizon URL using command line? Any command or config file exists? Or maybe via Fuel REST api somehow?
12:06 fatdragon joined #fuel
12:12 mskalski ilozgachev: horizon dashboard should be present on public vip, you can get ip address by executing: hiera public_vip on any openstack nodes but maybe there is a better way
12:13 degorenko joined #fuel
12:14 iberezovskiy joined #fuel
12:14 degorenko joined #fuel
12:15 skolekonov joined #fuel
12:23 mwhahaha_ joined #fuel
12:25 asaprykin joined #fuel
12:39 iberezovskiy joined #fuel
12:39 skolekonov joined #fuel
12:45 Verilium joined #fuel
12:45 iberezovskiy joined #fuel
12:52 code-R joined #fuel
13:39 GMAzrael joined #fuel
13:40 dguryanov2 joined #fuel
13:45 jobewan joined #fuel
13:49 code-R_ joined #fuel
13:52 code-R_ joined #fuel
14:07 fatdragon joined #fuel
14:08 Julien-zte joined #fuel
14:12 code-R joined #fuel
14:19 hyperscsi joined #fuel
14:20 hyperscsi If I'm trying to get a new fuel master node stood up, what ports should i have my firewall guy open up? fuel 8.0 and im planning on creating a local repo
14:24 mwhahaha hyperscsi: you mean to the outside world?
14:25 hyperscsi yes
14:26 mwhahaha if it's just ports, ntp, dns, http/https, rsync
14:26 mwhahaha whatever those standard ports are
14:29 fatdragon joined #fuel
14:29 hyperscsi ok thanks
14:50 hyperscsi if fuelmenu is not working, is astute.yaml the only file I need to edit with the settings?
14:56 xarses_ joined #fuel
15:03 fatdragon joined #fuel
15:04 mwhahaha hyperscsi: no, what's fuelmenu doing?
15:04 hyperscsi stuck on "applying changes..." when trying to save and quit
15:05 mwhahaha does it not yet have internet connectivity?
15:06 hyperscsi it wont until i configure fuelmenu to point to our internal dns servers
15:07 mwhahaha wonder if it's hanging on that
15:08 hyperscsi the last time I tried to deploy this it didnt have internet access and fuelmenu worked
15:08 hyperscsi but that was on 6.1
15:13 mwhahaha 6.1 didn't have the hard requirement on internet like 7+ does
15:13 krypto joined #fuel
15:13 mwhahaha so it's probably doing something, you could switch over to another terminal and see what it's doing in the background
15:14 hyperscsi hm
15:16 krypto joined #fuel
15:20 krypto joined #fuel
15:21 code-R joined #fuel
15:29 pradiprwt joined #fuel
16:00 xarses Weekly meeting in #openstack-meeting-alt
16:11 Zer0Byte__ joined #fuel
16:24 raunak joined #fuel
17:03 GMAzrael Is it posible to RAID1 a boot drive for nodes?
17:06 Zer0Byte__ i dont see any problem for that GMAzrael
17:07 Zer0Byte__ you mean hardware raid?
17:07 GMAzrael I created the RAID1 in Intel's BIOS, but it is still being reported as seperate drives in Fuel
17:07 Sketch because it's software raid
17:07 Zer0Byte__ and the sata mode is RAID?
17:07 Zer0Byte__ or AHCI?
17:12 GMAzrael RAID
17:12 GMAzrael and created the virtual drive in the bios
17:13 vvalyavskiy left #fuel
17:23 mwhahaha fuel only picks up what's being reported to the OS
17:23 mwhahaha so if the drives are showing up as 2 different devices that's something outside of fuel's world
17:25 GMAzrael will fuel pick up /dev/md0?
17:25 GMAzrael or only /sdX
17:25 mwhahaha if it's md0 that's software raid you'd have to configure that with fuel
17:26 Sketch can you configure software raid with fuel?
17:26 Sketch well, i guess you can always do it the hard way, if not via the web
17:26 mwhahaha yea the hard way not via the ui
17:27 mwhahaha http://docs.openstack.org/developer/fuel-docs/devdocs/develop/nailgun/customization/partitions.html
17:27 Zer0Byte__ Mr. mwhahaha  how are you
17:27 mwhahaha it supports a 'raid' type
17:27 mwhahaha which is raid-1
17:27 mwhahaha Zer0Byte__: hi
17:27 mwhahaha brb
17:29 code-R joined #fuel
17:32 fatdragon joined #fuel
17:50 fatdragon joined #fuel
17:56 pchaigno joined #fuel
18:03 GMAzrael after setting the os volume to raid in openstack.yaml, how is it propogated to nailgun. Just a restart of the service?
18:04 xarses you have to dump the database and reload it to re-read openstack.yaml
18:04 xarses you should be able to make a similar edit in the data that is given back from the release api, and upload that back
18:05 xarses then you will need to create a new cluster
18:08 GMAzrael not seeing a "dump db" command
18:08 failart joined #fuel
18:09 GMAzrael FYI - I'm sysadmin, not a developer
18:11 xarses after you stop nailgun/assassin/collect? you can `manage.py dropdb && manage.py syncdb && manage.py loaddefault` keep in mind, that it will literally drop everything in the database and no preserve anything
18:12 xarses s/no /not
18:12 GMAzrael not worried about it
18:14 GMAzrael thank you for the knowledge
18:50 hyperscsi is there a way to specify that my eth1 interface is for the gui and not eth0?
18:50 hyperscsi for fuel 8.0
18:52 krotscheck left #fuel
18:53 hyperscsi nevermind
18:54 hyperscsi for some reason my fuel master node isnt accepting ssh connections
18:54 hyperscsi but it was fine before when i set the eth0 interface as the "admin" interface
19:21 GMAzrael @xarses time to drop database?
19:22 xarses ?
19:22 GMAzrael I'm on a fresh 9.0 install and database drop is at 1hr and going
19:22 xarses oh, one of the processes is still running and holding the db lock hostage
19:23 GMAzrael I stopped nailgun, assassind, and mcollective
19:23 GMAzrael any others?
19:24 xarses `ps axu | grep nailgun` should tell you
19:24 xarses mcollective dosn't have to be stopped
19:24 xarses just the python processes
19:24 xarses there is a collect something process
19:25 GMAzrael postgress nailgun idle in transaction
19:25 xarses and maybe one other that I can't recall at the moment
19:25 * xarses starts a 9 environment
19:26 GMAzrael cobbler?
19:26 GMAzrael /usr/bin/ostf-server
19:26 xarses cobbler is fine
19:26 GMAzrael /usr/bin/statsenderd
19:27 xarses stat, yes
19:27 xarses ostf uses the db, but a different table
19:27 xarses database schema even
19:30 GMAzrael just restarted postgres and it worked
19:30 GMAzrael wierd
19:30 code-R joined #fuel
19:41 GMAzrael bad thing, Editing disks no longer works
19:42 GMAzrael Failed to generate volumes for node 'Untitled (14:6c)': ''file_system''
19:53 xarses something in the data you provided caused problems with the code that generates the volumes
19:53 xarses which then breaks the UI
19:54 raunak joined #fuel
20:03 GMAzrael just changing the OS type from vg to raid?
20:18 xarses possibly, it would depend on the inputs required by the function
20:18 xarses you can look to see how the /boot volume is set up as raid and follow that
20:18 xarses otherwise there may be a bug
20:39 GMAzrael I don't see an item in openstack.yaml for /boot
20:43 GMAzrael all I see is the root and swap partitions
21:00 xarses hmm, I can't find it either I swore it used to be there
21:00 xarses It might be some implicit magic with the os partition
21:03 AlexAvadanii :) yes, nailgun driver adds /boot silently, considering raid partitions for example
21:04 AlexAvadanii https://github.com/openstack/fuel-agent/blob/master/fuel_agent/drivers/nailgun.py#L139
21:06 xarses GMAzrael: are you just trying to raid1 /boot?
21:06 GMAzrael raid1 OS drives
21:07 GMAzrael mirrored 80GB SSD
21:09 GMAzrael to prevent having to rebuild the entire storage node should my OS drive fail
21:09 hyperscsi I just finished installed fuel 8.0 on my master node. I applied maintenance updates but for some reason I cannot ssh into the fuel master. it is refusing the connection. any ideas?
21:10 hyperscsi tried rebooting as well but still refusing connection
21:11 mwhahaha probably an iptables rule
21:11 AlexAvadanii hyperscsi: using 10.20.0.2 or a public interface? ssh only listens on 10.20.0.2 in fuel 8.0
21:11 AlexAvadanii mwhahaha: nah, that is fuel 9.0 :P
21:11 mwhahaha pretty sure it's also 8
21:11 mwhahaha where if you switch from eth0 it locks you out
21:11 AlexAvadanii we never had issues with iptables and ssh in 8, only in 9; in 8 we had to change the listening address for sshd though
21:12 mwhahaha i would login via console and check it out
21:12 hyperscsi yeah i changed defaults. eth0 is my interface with an IP i can reach insdie my network, eth1 im using for PXE and is a private address only local to the switch where all my nodes are connected
21:13 GMAzrael am I off base on my recovering thinking?
21:14 hyperscsi ill look into that for sshd, if you have some instructions that would be great. thanks AlexAvadanii and mwhahaha
21:14 fatdragon joined #fuel
21:14 AlexAvadanii hyperscsi: mwhahaha is probably right then :) for iptables `iptables -A INPUT -p tcp --dport ssh -j ACCEPT`, for sshd, edit /etc/ssh/sshd_config & service sshd restart (or something along these lines)
21:15 hyperscsi awesome thanks!!! it was sshd
21:16 hyperscsi was only listening on the address assigned to eth1
21:20 GMAzrael md126 "should" be presented as an option for the OS to install on correct?
21:38 AlexAvadanii should is an ambiguos level of expectation ... my understanding is that it should, but maybe some printf debugging would help :)
21:39 AlexAvadanii when I had zero understanding/knowledge of the nailgun part of fuel, I often added print statements across fuel_agent/drivers/nailgun.py, restarted nailgun service and checked the logs, it usually helps troubleshooting wihout having to really understand the whole code
21:41 fatdragon joined #fuel
21:45 AlexAvadanii1 joined #fuel
21:47 Egyptian[Home] joined #fuel
21:51 GMAzrael Alex, as a SysAdmin, can you elaborate on the printf?
21:53 Egyptian[Home] joined #fuel
21:53 AlexAvadanii1 GMAzrael: it's a low level method of debugging I use when I run out of logical approaches; basically I start with adding random blocks of 'print "checkpoint 1"` in the nailgun.py file for example, inside functions that look related to what you are doing
21:54 AlexAvadanii1 then, `service nailgun restart`, try doing you work as you would normally do in the web ui, and check the logs section in fuel web ui
21:54 GMAzrael lack of knowledge on the entire code of nailgun presents a challenge of that.
21:55 AlexAvadanii1 well, print is the generic python printing instr, in this case you would want something like LOG.debug('bla.')
21:56 GMAzrael Our IT group just went to Agile task management, and I don't believe I have the hours to spend debugging nailgun to find why mdXXX isnt presented as an option.
21:56 AlexAvadanii1 GMAzrael: I knew nothing when I started, and I still get lost tracking the complete flow, grep is your friend, you can start by looking for the error you get + debug logs from the web ui page (.e.g https://<fuel ip>:8443/#cluster/1/logs/type:local;source:app;level:debug)
21:57 GMAzrael Thanks Alex
21:58 AlexAvadanii1 maybe try dumping everything you can and pastebin it, and someone else might have a bright idea just by looking at `lspci`, `lshw` or whatever else involved in that (like `lsblk`)
21:58 GMAzrael I can see the device in mdadm --detail mdXXX
21:58 GMAzrael fuel just doesnt present it as an option to install the OS on
22:00 AlexAvadanii1 mwhahaha: you usually were on point regarding fuel-agent stuff, no ideas?
22:04 Egyptian[Home] joined #fuel
22:04 GMAzrael before I changed the openstack.yaml file, there was no error generated
22:05 mwhahaha not really, this one isn't in my area of expertise.  i just recall that raid is support somehow but not sure how to do it
22:05 mwhahaha GMAzrael: are you not using ceph for your storage node?
22:05 mwhahaha personally i think software raid is more trouble than it's worth
22:06 GMAzrael I am using ceph for all my storage, I'm talking about the OS install on the slaves
22:06 mwhahaha what slaves
22:06 GMAzrael Compute/storage/controller
22:06 AlexAvadanii1 +1 on software raid being a pain :) mdadm has some serious issues, at least the one we use in Ubuntu
22:06 mwhahaha the whole point is to allow for failure via multiple controllers, ceph for storage, compute should be throwaway
22:07 GMAzrael if the OS drive fails for a storage node, don't all the ceph OSDs have to rebuild when the system comes back online?
22:08 GMAzrael or does it just reprovision the OS and keep the OSD's?
22:08 mwhahaha if you have your replication factor set correctly you won't end up with any data loss and things should continue but yes there would be some impact when they come back on line but that should be relatively minimal
22:11 GMAzrael just worry about the rebuild time and bandwhith requirements for 48TB
22:11 mwhahaha then buy something that does hardware raid ;)
22:12 GMAzrael oh goody
22:13 GMAzrael funny thing is this is the same hardware as the unlocked appliance
22:13 mwhahaha yea but they do provisioning differently
22:15 * mwhahaha wanders off
22:22 AlexAvadanii1 GMAzrael: since no error was generated before editing openstack.yaml, maybe something is not right there? https://docs.fuel-infra.org/fuel-dev/develop/nailgun/customization/partitions.html says to make sure "our logical volume name (“LOGICAL_VOLUME_NAME” in the example above) is not the same as the volume group name (“os”)" [..]
23:00 Damjanek joined #fuel
23:56 xarses joined #fuel

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