Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-02-07

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

All times shown according to UTC.

Time Nick Message
00:49 IlyaE joined #fuel
01:06 jouston__ joined #fuel
01:57 jouston joined #fuel
02:31 jouston_ joined #fuel
03:02 rmoe joined #fuel
04:10 xarses joined #fuel
04:51 IlyaE joined #fuel
04:54 IlyaE joined #fuel
05:06 jouston joined #fuel
05:08 saju_m joined #fuel
06:45 mihgen joined #fuel
07:31 jouston joined #fuel
07:32 saju_m joined #fuel
07:41 IlyaE joined #fuel
08:18 jouston joined #fuel
08:40 rvyalov joined #fuel
08:51 miguitas joined #fuel
09:07 vk joined #fuel
09:55 mrasskazov joined #fuel
10:57 e0ne joined #fuel
11:01 jouston_ joined #fuel
12:39 jouston joined #fuel
12:45 Dr_Drache joined #fuel
12:46 Dr_Drache xarses, so, I might not be able to continue using fuel.
12:46 Dr_Drache have to see how the "your not invited" meeting goes
12:52 mattymo Dr_Drache, what meeting?
12:52 Dr_Drache mattymo, of the CFO/and my director of IT.
12:54 Dr_Drache rather PITA to have someone watch over who has little technical experence.
12:57 mattymo it does take some time to learn how OpenStack works. When Fuel meets all your needs, yes, it's easy, but it still doesn't show you all the relationships until you start diggging
13:00 Dr_Drache fuel meets the needs of deployment, but with the percived lag between what is said from marketing is in it, and what acually works, is a huge gap. and people are not fond of that.
13:02 mattymo which inaccurate statements are you referring to?
13:02 Dr_Drache i'd have to go over the transcripts of our phone meetings.
13:03 Dr_Drache to me, it's not THAT big of a deal, I'm here getting things fixed for more people than me.
13:04 mattymo Okay. Let me know if I can help
13:04 TVR__ joined #fuel
13:06 Dr_Drache xarses, and MiroslavAnashkin have been doing a great job, I've had some bug fixes, and sorting things out, of course when a bug shows up and the dev says "you should really upgrade, that's a VERY VERY old version" that's a bit annoying :P
13:06 TVR__ morning folks
13:07 Dr_Drache TVR__, morning... great news. there is a meeting that i'm not invited to, and it's about openstack/fuel, LOL
13:07 TVR__ so.. adding ssh keys:
13:07 TVR__ I go to access and security
13:07 TVR__ create a pem file
13:07 TVR__ download it
13:07 TVR__ put it on the server to connect
13:07 TVR__ chmod 600 it
13:08 TVR__ ssh-add it
13:08 TVR__ add a floating ip to instance that uses it
13:08 TVR__ try so ssh and get challenged with a user / password
13:08 TVR__ what am I doing wrong?
13:09 TVR__ quit my job and move to windows?
13:09 TVR__ heh
13:09 Dr_Drache yes.
13:09 Dr_Drache lol
13:12 Dr_Drache TVR__, seems correct.
13:12 Dr_Drache I can't try, because my cluster is dead to the world.
13:13 TVR__ this should work for.. root @ instance.. yes? did I forget to attach the key to root somehow?
13:14 Dr_Drache it shouldn't matter where you attached it client side, as long as it's attached to the invoking use.
13:14 Dr_Drache user.
13:15 Dr_Drache hope MiroslavAnashkin is around soon, want to see if he found out anything about why ubuntu doesn't install.
13:15 Dr_Drache well, boot after install
13:16 TVR__ yea.. so as I am setting parameters for building the instance.. I go to Access & Security tab and associate the key to it.. I don't need to set admin user and pass, right?
13:16 TVR__ or dont need to set admin pass.. right?
13:17 Dr_Drache for the instance, i don't think so. but honestly, I haven't been able to get that far to know for SURE.
13:18 Dr_Drache since my node died.
13:18 Dr_Drache (s)
13:19 TVR__ ok.. will wait for the guys then...they probably were celebrating after their skating pair simply annihilated all others in the Olympics last night... it was greater than 10 point difference... as they were flawless...
13:19 Dr_Drache damn
13:19 Dr_Drache not bad
13:21 TVR__ yea.. I don't pay much attention to the skating pairs, but when it was that much of a difference and they simply made all the others look like they were at a high school dance.. well, you notice... I like the snowboarding  and skiing bits myself... USA kid cracked a rib and still did well.. but he has to still do a few more runs to not be eliminate
13:21 TVR__ d...
13:22 TVR__ I guess the snowboarding freestyle is a bit more than they have ever put out before..
13:26 mattymo Dr_Drache, do you have network issues with ubuntu install in Fuel 4.0?
13:26 Dr_Drache mattymo, nope, I have "won't boot"
13:26 mattymo oh cool
13:27 Dr_Drache never gets far enough to know.
13:27 Dr_Drache well, on the production hardware anyway.
13:27 mattymo so it may be related to https://bugs.launchpad.net/fuel/+bug/1269920
13:27 mattymo there are 2 patches attached to it
13:27 mattymo we haven't released a bugfix iso unfortunately
13:28 Dr_Drache older HP works "fine" but, new dells freeze on grub. (can't detect drives)
13:31 mattymo Dr_Drache, would you be kind enough to share a ticket # with support or file a bug at bugs.launchpad.net/fuel?
13:32 mattymo if bootstrap booted fine on it, it should have enough details in the diagnostic snapshot to identify the disk controllers
13:32 Dr_Drache MiroslavAnashkin, has my logs and information about it. centOS installs and boots fine.
13:33 Dr_Drache I recall him asking me to wait to see if i need to file a bugreport on it.
13:36 Dr_Drache I recall him saying he'd talk to.... someone who started with an A. to see if I need to file it.
13:37 Dr_Drache it installs perfectly according to the snapshot, just never boots.
13:37 mattymo there are like 8 guys in support whose name starts with A
13:37 mattymo :)
13:37 TVR__ look for the A+ as thats a better grade?
13:38 Dr_Drache mattymo, I relize that when I said it, but it was a "russian" sounding name. best I can do right this min
13:39 TVR__ so.. here is a question.. when associating keypairs when creating an instance... it is supposed to create a .ssh directory and populate it.. in the instance.. yes?
13:39 mattymo TVR__, yes in /root/.ssh/authorized_keys2
13:39 TVR__ ok.. does .ssh directory need to already exist?
13:39 mattymo it's part of cloud-init, if I remember correctly
13:41 anotchenko joined #fuel
13:41 TVR__ my issue might be that my image I spun does not have a .ssh directory...
13:41 mattymo sorry, but I'm not 100% sure on this one
13:41 TVR__ ok.. cool
13:45 Dr_Drache might as well do VLANs today
14:16 jaypipes joined #fuel
14:17 Dr_Drache ummm
14:17 Dr_Drache TVR__,
14:17 TVR__ ?
14:17 Dr_Drache seen this : Untitled (0B:A0)d4:ae:52:ea:0b:a2eth11024,1025,1026,1027,1028,1029,1030,1000,1001,1002,1003,1004,1005,1006,1007,1008,1009,1010,1011,1012,1013,1014,1015,1016,1017,1018,1019,1020,1021,1022,1023
14:18 TVR__ no.. what log?
14:18 Dr_Drache http://paste.openstack.org/show/63049/
14:18 Dr_Drache I did the fuel test.
14:19 Dr_Drache those interfaces are untagged networks.
14:20 TVR__ the defaults are VLANs 1000 to 1030 which is what it is looking for
14:20 TVR__ if you had not edited anything
14:20 Dr_Drache so, i need to tag those ports with those vlans?
14:21 Dr_Drache in the config there is no option to touch tagging for the Neutron L3 Tagging.
14:24 Dr_Drache so confusing... LOL
14:25 Dr_Drache not really, just newer at this vlan stuff than i'd like to be
14:33 Dr_Drache seems my issue is about the same as https://bugs.launchpad.net/fuel/+bug/1249307
14:33 Dr_Drache that's un-reproduced
14:35 Dr_Drache shit
14:35 Dr_Drache i'm an idiot.
14:37 mattymo if your network verification fails, it means your network config is physically not correct (or switch conf)
14:38 Dr_Drache mattymo, i'm an idiot, I glossed over the Neutron L2 vlan range.
14:38 Dr_Drache don't understand what is required there at this time, but that's a user error
14:39 TVR__ everything works for me network wise.. but my verification test fails with Verification failed.
14:39 TVR__ Network verification on Neutron is not implemented yet
14:40 TVR__ I think you guys have a bug on that though.. or haven't implemented all the test suits yet for neutron
14:40 Dr_Drache ahh, TVR__ the l2 vlan range, happen to know in idiot terms what that is for?
14:41 TVR__ the guy who was on yesterday can explain far better than I.. I ~get~ networking.. but don't always know what or why I get it
14:42 Dr_Drache same.
14:42 Dr_Drache all good.
14:43 Dr_Drache will deploy and ignore those test errors
14:51 anotchenko joined #fuel
14:56 jkirnosova joined #fuel
15:20 Dr_Drache nope
15:21 MiroslavAnashkin Dr_Drache: Please try this initramfs.img It is from 4.1 and includes all patches. https://drive.google.com/file/d/0B4szgxpvz2aCUVVDbkZIRmxtQjQ/edit?usp=sharing
15:22 MiroslavAnashkin Dr_Drache: We hope it should solve ubuntu installation
15:22 Dr_Drache ok
15:25 Dr_Drache MiroslavAnashkin, happen to have the link for the hardlinks?
15:30 MiroslavAnashkin Dr_Drache: Yes
15:31 MiroslavAnashkin Dr_Drache: Hmm, please do not use this image. One more critical patch just came out
15:31 Dr_Drache ok, lOL
15:33 ogelbukh joined #fuel
15:33 IlyaE joined #fuel
15:33 Dr_Drache I seem to not be getting the setup for the vlans
15:34 TVR__ root
15:34 TVR__ wrong window
15:35 vk joined #fuel
15:36 jouston joined #fuel
15:36 Dr_Drache lol
16:01 TVR__ ok.. so anyone want to field a ssh keys issue? I installed cloud-init on my qcow2 image... but when it is booting, it trieds to connect bach to 169.254.169.254 and gets no route to host
16:02 TVR__ that connection initiates from the DataSourceEc2.py
16:03 TVR__ someone please smack me upside the head with what I overlooked please..
16:12 TVR__ NOZEROCONF=yes in /etc/sysconfig/network seems to fix the route issue
16:24 TVR__ so the instance gets renamed now...
16:25 TVR__ but the authorized_keys never gets created
16:25 Dr_Drache ahh, that's a crappy bug
16:26 TVR__ I am probably doing something wrong here... but I followed the instructions from http://docs.openstack.org/user-guide/content/keypair_add.html and http://docs.openstack.org/image-guide/content/centos-image.html
16:43 anotchenko joined #fuel
17:05 xarses joined #fuel
17:18 mihgen joined #fuel
17:19 MiroslavAnashkin joined #fuel
17:30 TVR__ so I had to add some code to rc.local as the cloud-init wasn't doing what it was supposed to (in centos) but it works now
17:31 TVR__ Dr_Drache ... you will probably see this as well.. so ping me if you have issue with ssh keys and I will pass you my notes
17:36 TVR__ now adding capacity and seeing if anything breaks
17:41 angdraug joined #fuel
17:41 Dr_Drache TVR__, I'm still trying to sort out vlans.
17:41 TVR__ I gave up on VLANs for now.. gre for the time being
17:42 Dr_Drache I don't understand why it is looking for 1001-1030 for untagged network
17:44 Dr_Drache i'm waiting right now on MiroslavAnashkin, see if that critial patch is going to be posted
17:49 Dr_Drache that, and I broke fuel
17:50 TVR__ ok.. my keys work.. yay...
17:50 MiroslavAnashkin Dr_Drache: Vkozhukalov postponed the patch to the Monday.
17:52 Dr_Drache ok
17:52 Dr_Drache can I see the bug report and see if that would effect me?
17:56 MiroslavAnashkin Dr_Drache: There is Neutron L2 Configuration section on Networks tab. You may configire VLAN numbers there. VLANs are necessary for Neutron+VLAN segmentation config to sort out the traffic between different Openstack tenants. Every tenant require one personal VLAN
17:58 MiroslavAnashkin In turn, tenant may include a full set of virtual network inftastructure - routers, bridges, ports, connections etc. And it even uses its own internal VLANs
18:03 vk joined #fuel
18:15 Dr_Drache MiroslavAnashkin, understandable.
18:15 Dr_Drache on the phyiscal switch though
18:15 Dr_Drache what needs to be tagged?
18:16 rvyalov joined #fuel
18:17 MiroslavAnashkin These VLANs, configured in L2 segment section plus all VLAN numbers near each "Use VLAN tagging" check box, if checkbox set to ON
18:18 Dr_Drache so, default, it says vlans 1001 to 1030, I need to have all my ports tagged with those access vlan
18:18 Dr_Drache PLUS, "use vlan tagging" tagged
18:19 MiroslavAnashkin Yes, while Openstack start to use these VLANs 1001 to 1030 one by one with every new tenant creation
18:20 Dr_Drache that's alot of vlans to create on switches
18:23 MiroslavAnashkin 30 is very reduced number. We used 1000 by default, but network verification of 1000 VLANs exceeds FUEL timeouts. It take about hour. So, we reduced default number until we rewrite verification.
18:24 e0ne joined #fuel
18:25 TVR__ wow.. every single instance gets put on it's own vlan....So, in projects, they each still all get their own individual vlan rather than just all be on the same vlan?
18:26 Dr_Drache I don't see how vlans make anything simpler.
18:26 Dr_Drache lol
18:27 Dr_Drache since you need to have a vlan setup and working for each one, add alother switch, you just doubled the work.
18:27 TVR__ if every instance in a project is on the same vlan, that will isolate all their traffic and that would be a good thing... so I can see it being very helpfull
18:28 Dr_Drache I'm only looking at it from the admin/setup side.
18:28 Dr_Drache not the "once running" side.
18:28 Dr_Drache lol
18:29 MiroslavAnashkin Every instance is assigned its personal internal VLAN, which completely exists inside compute node only in between the instance (br-int, br-ex) and br-ethX. Neutron changes this internal VLAN tag on the br-ethX to the proper tenant VLAN number (one of the L2 segment VLAN pool)
18:29 Dr_Drache oh
18:30 Dr_Drache so, those vlans don't exist outside software.
18:30 MiroslavAnashkin All these VLANs exist inside OpenStack.
18:31 MiroslavAnashkin But since OpenStack uses real network hardware  - these VLANs should be enabled on this hardware.
18:32 Dr_Drache so i was correct?
18:35 mattymo joined #fuel
18:36 Dr_Drache really, I wanted to be wrong.
19:05 TVR__ the ceph install for this should be able to take swift and S3 commands, yes? RadosGW is all set up if you check that in settings and such.. yes?
19:06 TVR__ I keep seeing WARNING: cannot read region map
19:06 TVR__ and my swift calls end with [Errno 111] Connection refused
19:22 Dr_Drache MiroslavAnashkin, is it "safe" to test with that initramfs.img?
19:25 MiroslavAnashkin No. it reports incorrect data and so, you get red screen on node reboot
19:25 Dr_Drache ahh bummer
19:25 Dr_Drache wash hoping to deploy and test on buntu.
19:25 Dr_Drache all is good, hopfully ceph doesn't die this time
19:40 MiroslavAnashkin joined #fuel
19:46 TVR__ the ceph object store installed by fuel is the RadosGW .. correct?
19:50 MiroslavAnashkin joined #fuel
20:06 TVR__ I see it on node-1 my primary node... ceph-radosgw is started.... and a restart goes well....
20:16 TVR__ my issue is.... it looks like it is set up the same as my home-rolled ceph cluster...but I can connect to my ceph cluster API, but I don't connected to the openstack ceph
20:17 IlyaE joined #fuel
20:17 TVR__ not many logs to look at
20:19 TVR__ [root@controller ~]# swift -V 1.0 -A http://ceph01.athenalab.athenahealth.com/auth -U fsmith:swift -K GeDhs8s4v35AeEt0SYsuT+GTZMHnQWbk+Pj3Xr2k stat
20:19 TVR__ Account: v1
20:19 TVR__ Containers: 1
20:19 TVR__ Objects: 2
20:19 TVR__ Bytes: 3446752
20:19 TVR__ Server: Apache/2.2.15 (Oracle)
20:19 TVR__ Connection: close
20:19 TVR__ X-Account-Bytes-Used-Actual: 3452928
20:19 TVR__ Content-Type: text/plain; charset=utf-8
20:19 TVR__ [root@controller ~]# swift -V 1.0 -A http://node-1.athenalab.athenahealth.com/auth:6780 -U fsmith:swift -K "EAjVkOxZqkSpRp0ute2K1JWi5R44nl261Fv\/EGDN" stat
20:19 TVR__ [Errno 111] Connection refused
20:21 TVR__ [root@controller ~]# swift -V 1.0 -A http://node-1.athenalab.athenahealth.com/auth -U fsmith:swift -K "EAjVkOxZqkSpRp0ute2K1JWi5R44nl261Fv\/EGDN" stat
20:21 TVR__ [Errno 111] Connection refused
20:22 TVR__ any thoughts? any kicks in the right direction?
20:38 Dr_Drache let me see
20:38 Dr_Drache was in meeting
20:38 Dr_Drache for sure it looks like it's either blocked, or disabled
20:54 TVR__ I filed a ticket ... they are very responsive...
20:54 TVR__ nothing jumps out at me yet
20:55 TVR__ 11   ACCEPT     tcp  --  0.0.0.0/0            0.0.0.0/0           multiport dports 6780 /* 012 RadosGW allow */
20:55 TVR__ that suggests it is not the iptables firewall
20:56 Dr_Drache i'll start real support once I can deploy in ubuntu.
20:57 TVR__ swift -V 1.0 -A http://node-1.athenalab.athenahealth.com:6780/auth -U fsmith:swift -K EAjVkOxZqkSpRp0ute2K1JWi5R44nl261Fv\/EGDN stat
20:57 TVR__ Account: v1
20:57 TVR__ Containers: 0
20:57 TVR__ Objects: 0
20:57 TVR__ Bytes: 0
20:57 TVR__ Server: Apache/2.2.15 (CentOS)
20:57 TVR__ Connection: close
20:57 TVR__ X-Account-Bytes-Used-Actual: 0
20:57 TVR__ Content-Type: text/plain; charset=utf-8
20:57 TVR__ DOHHHHH !!!!!
20:57 TVR__ :6780/auth ... not /auth:6780
20:59 TVR__ I just think it is about that time and I am fried
21:00 TVR__ rookie mistake... glad I caught it and didn't waste anyones time
21:00 Dr_Drache it's friday.
21:00 Dr_Drache question.
21:01 Dr_Drache where are you setting your TLD?
21:02 Dr_Drache manually?
21:10 TVR__ TLD?
21:11 TVR__ top level domain?
21:12 TVR__ I just (for node-1) set it in hosts for the time being..
21:17 jouston_ joined #fuel
21:51 mutex ug
21:52 mutex neutron just is not reliable for me
21:53 IlyaE joined #fuel
22:25 vk joined #fuel
22:32 dhblaz joined #fuel
22:34 dhblaz I can't seem to get the "Customization Script" found in the Launch Instance feature of the dashboard to work.  Even if my instance contains cloud-init.  I see this in the console log:
22:34 dhblaz http://paste.openstack.org/show/63134/
22:49 jouston_ joined #fuel
23:11 dhblaz It seems that the neutron "routing" to the metadata service is fragile
23:11 dhblaz I can make it work some times on some instances
23:11 dhblaz but not reliably
23:15 richardkiene joined #fuel
23:24 richardkiene_ joined #fuel
23:29 richardkiene_ joined #fuel
23:46 e0ne joined #fuel
23:55 richardkiene joined #fuel

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