Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-08-22

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

All times shown according to UTC.

Time Nick Message
00:13 KodiakF joined #fuel
00:17 mattgriffin joined #fuel
00:27 kupo24z xarses: when is the corosync network verification need to be checked?
00:27 kupo24z does*
00:28 xarses it doesn't right and will likely report false failures so, don't worry about it untill 6.0
00:28 xarses I'll file a bug about it, i thought it was hidden from the ui
00:29 kupo24z Is the zabbix support currently working?
00:29 xarses It's supposed to be
00:30 kupo24z kk ill try it out
00:30 rmoe joined #fuel
00:30 xarses can you send me the /api/version data?
00:31 kupo24z {"build_id": "2014-08-21_23-01-26", "ostf_sha": "907f25f8fad39b177bf6a66fba9785afa7dd8008", "build_number": "85", "auth_required": true, "api": "1.0", "nailgun_sha": "e9dd053fce12908b5a5a404435ce2e8200450a23", "production": "docker", "fuelmain_sha": "5a7df58786db7962a1774b5be3611c4e7543015d", "astute_sha": "ac520b09525af4551e730b1c1f78170fefaf3cb8", "feature_groups": ["experimental"], "release": "5.1", "fuellib_sha": "bddba1e854a6b0350e8
00:32 tatyana joined #fuel
00:32 xarses kupo24z: https://bugs.launchpad.net/fuel/+bug/1360018
00:33 rmoe joined #fuel
00:34 kupo24z Cool
00:37 kupo24z Also found a new ceph problem, ill verify when this deployment is done though
00:37 kupo24z Something about the number of PG's in use was imbalanced on compute
00:58 xarses hmm
00:58 xarses let us know
00:59 xarses I'm sure angdraug would love another ceph bug
01:02 angdraug kupo24z:  not the same bug I fixed yesterday about pgp_num not being set?
01:07 kupo24z angdraug: Nope I think its different
01:08 kupo24z It was after I fixed that myself and started writing stuff to VM's
01:08 kupo24z btw I am seeing about 20% slower read/writes with firefly, but I think we are going to take that up with inktank
01:10 kupo24z Only way I saw any improvement over .79 was using unsafe
01:10 angdraug yeah, I'm sure sage will be thrilled :)
01:11 kupo24z Currently waiting on the dc guys to rack the 40~ storage nodes so I can test with that instead of this small cluster
01:14 mattgriffin joined #fuel
01:33 xarses joined #fuel
02:18 rmoe joined #fuel
03:14 mattgriffin joined #fuel
03:17 rmoe joined #fuel
04:05 jaypipes joined #fuel
04:09 ArminderS joined #fuel
04:16 rmoe joined #fuel
04:40 tatyana joined #fuel
05:28 HeOS joined #fuel
05:50 nimion joined #fuel
05:52 tatyana joined #fuel
06:05 Havard evg: are you in today? :-)
07:04 pasquier-s joined #fuel
07:10 hyperbaba joined #fuel
07:24 aepifanov joined #fuel
07:28 youellet joined #fuel
07:35 HeOS joined #fuel
07:46 monester_ joined #fuel
07:59 youellet joined #fuel
08:02 kaliya Havard: please try to ask
08:03 Havard kaliya: did you see my outputs yesterday from both the "cloud" server with fuel and my container?
08:03 Havard on the server i got requests on dhcp from the node, but not in the container
08:03 Havard so i wonder if there is some issue with my installation
08:05 kaliya yes, how to reproduce? did you just install fuel 5.0.1 or performed fuelmenu_
08:06 kaliya ?
08:06 Havard i did do the fuel setup by setting showmenu=yes on the server (to make it avaliable on public IP and to make the DHCP/PXE use the range of IPs provided by my CTO)
08:07 Havard as i do not have local access to the fuel/cloud server i need it to be available on a public IP
08:07 kaliya on metal or virtual setup?
08:08 Havard in the installation i did not get any options for metal or virtual
08:08 Havard i just completed the installation guide, and than tried to boot the nodes from pxe, but that did not work :-/
08:08 kaliya did you install on a hardware or in a virtual environment (virtualbox, esxi?)
08:09 Havard ah, this is installed on hardware
08:09 Havard the cloud/fuel server is a physical server
08:09 kaliya I've seen you did some tcpdump, I had a very similar problem when the Fuel ISO didn't finish the installation properly and containers were not ready somehow
08:09 Havard and the nodes (three of them) as well
08:09 Havard i just followed this guide: http://docs.mirantis.com/openstack/fuel/fuel-5.0/pdf/Mirantis-OpenStack-5.0.1-UserGuide.pdf
08:09 Havard and when i get to page 16 to boot the nodes nothing happens
08:09 kaliya that's the right guide, yes
08:10 Havard on the console of the nodes they do not find the dhcp server
08:10 kaliya so if on master you try to `dockerctl check` do you see some error?
08:10 Havard that is why we tried those tcpdumps yesterday
08:10 Havard moment
08:11 Havard http://pastebin.com/gk8CrDc1
08:11 Havard that is a dump of that command
08:13 Havard so evg asked me to do those tcpdumps, first on the fuel/cloud server, than in the cobbler container, and on the fuel/cloud it shows connection, but not in cobbler
08:14 kaliya could you please pastebin your /etc/fuel/astute.yaml ?
08:16 Havard i think i can see the error, but i will paste for you:
08:16 Havard http://pastebin.com/JwT1WRx7
08:16 Havard cidr: 10.20.0.0/24   <- is not in the same subnet
08:21 tuvenen joined #fuel
08:23 Longgeek joined #fuel
08:24 eshumkher joined #fuel
08:35 Longgeek joined #fuel
08:45 e0ne joined #fuel
08:52 monester_ joined #fuel
08:56 kaliya so Havard please try to run fuelmenu and set the correct network settings according to your physical configuration, then we retry :)
09:03 Havard kaliya: thanks, but that cidr is not in the fuelmenu anywhere
09:03 Havard strange
09:03 e0ne joined #fuel
09:38 adanin joined #fuel
09:41 evg Havard: hello
09:51 lordd_ joined #fuel
10:53 nimissa left #fuel
11:08 e0ne joined #fuel
11:21 e0ne joined #fuel
11:25 omelchek joined #fuel
11:43 Havard kaliya: still after running the fuelmenu the cidr is pointing to 10.20.0.0/24
11:44 Havard i am not sure if that is the cause of our problem, but it is for sure not pointing correct
11:49 Havard so still the same from the tcpdump as yesterday: http://pastebin.com/0pZMG6i4
11:56 pasquier-s joined #fuel
12:07 evg Havard: please show    "tcpdump -n -i eth1 port 67 or port 68"
12:07 Havard evg: from the container or from the fuel/cloud server?
12:16 evg Havard: from fuel node
12:17 evg Havard: just do you see packets's flow in two directions?
12:20 Havard evg: no :-(
12:20 Havard http://pastebin.com/zFBwJwa3
12:20 HeOS joined #fuel
12:24 omelchek joined #fuel
12:24 lordd_ joined #fuel
12:27 evg Havard: beg your pardon: "tcpdump -n -i eth1 port 68 or port 69"
12:27 Havard ah, i will try 69 now
12:27 evg Havard: ok
12:27 lordd_ joined #fuel
12:27 Havard :-)
12:27 Havard thanks
12:27 e0ne joined #fuel
12:31 Havard evg: same on port 69, no traffic
12:36 evg Havard: aah, my mistake again, i mean you run "tcpdump  -n -i eth1 port 68" on fuel master and then reboot your node
12:36 evg Havard: you should see dhcp traffic in the both directions
12:36 Havard yeah, i did run that and rebooted my node
12:37 Havard only the first command from yesterday gave some kind of respond on the tcpdump, all others give no results
12:40 evg Havard: then try "tcpdump  -n -i eth0 port 68" and reboot the node
12:41 evg Havard: so you've reconfigured your env and today we're debbuging another case?
12:42 Havard i did the fuelmenu again today, but the case is the same as yesterday
12:43 Havard the only thing i have found so far is this cidr  pointing to another range than in my setup, but that has not been changed
12:43 Havard i still have the same problem that my nodes can not get IP from the fuel/cloud server
12:44 Havard eth0 does not give any respond either
12:44 Havard but my network is on eth1 for the dhcp and pxe:
12:44 Havard eth1      Link encap:Ethernet  HWaddr 00:25:90:92:8E:41
12:44 Havard inet addr:10.217.180.5  Bcast:10.217.180.255  Mask:255.255.255.0
12:44 Havard eth0 is the public IP for the server
12:46 dilyin joined #fuel
12:55 artem_panchenko Havard: hi. Show an output of this command on fuel server: "ps aux | grep dhcrela[y]" Possibly dhcrelay is listening on eth0 instead of eth1
12:56 evg right
12:57 Havard [root@cloud ~]# ps aux | grep dhcrela[y]
12:57 Havard root      3876  0.0  0.0 108164  1456 ?        S    11:53   0:00 /bin/bash /usr/local/bin/dhcrelay_monitor
12:57 Havard root      3942  0.0  0.0   6648   296 ?        Ss   11:53   0:00 /usr/sbin/dhcrelay -i eth1 -i docker0 172.17.0.3
12:57 evg Havard: and please show your /etc/fuel/astute.yaml
12:58 Havard http://pastebin.com/BSXhzSBJ
12:58 artem_panchenko Havard: let's check that dhcp-server is working in container, please run 'dhcpcheck discover --ifaces eth1' on fuel server and provide its output
13:02 Havard give me a few minutes, on a support call with a client :-)
13:12 Havard http://pastebin.com/9YVLiJUq
13:13 artem_panchenko Havard: execute 'export LC_ALL=en_US.utf8' first and try again
13:14 Havard http://pastebin.com/Rm7G58Q9
13:18 artem_panchenko Havard: great, so DHCP server works correctly and dhcrelay is listening on eth1. I believe slaves can't get IP and boot over PXE because they are connected to the another L2 network segment than eth1 interface on Fuel server.
13:34 Havard hmm, strange
13:35 Havard i will have to find a cisco guy to solve that
13:35 nimion joined #fuel
13:37 Havard thanks a lot guys, i appreciate all help *thumbs up*
13:38 evg Havard: yes, let's' focus on network functioning, you have to see bootp traffic on eth1
13:38 Havard evg: i think that some of the switch ports might be tagget with a vlan
13:39 Havard evg: and that might be the reason why it is not working
13:39 Havard evg: i will try to get hold of my CTO during the evening to check the switch config
13:39 Havard evg: i know absolutely 0 about switch config :-)
13:40 Havard i do not even know how to login to them
13:46 mattgriffin joined #fuel
14:17 lordd_ joined #fuel
14:21 evg Havard: ok. we're awaiting news from you
14:33 lordd_ joined #fuel
15:13 mattgriffin joined #fuel
15:15 jobewan joined #fuel
15:16 DaveJ__ joined #fuel
15:27 xarses joined #fuel
15:39 blahRus joined #fuel
15:54 jobewan joined #fuel
15:56 adanin joined #fuel
16:19 geekinutah joined #fuel
16:30 rmoe joined #fuel
16:31 adanin joined #fuel
17:21 Longgeek joined #fuel
17:32 aepifanov joined #fuel
17:58 xarses joined #fuel
18:08 leakypipes joined #fuel
18:10 kupo24z joined #fuel
18:20 e0ne joined #fuel
18:20 ganso Hi guys! I see something weird in my horizon (and maybe this is related to my tempest quotas tests failing), in admin -> system info -> default quotas, I see -1 for everything, and in overview the pizza graphs do not work. When I had setup a SLES environment before everything was set automatically according to my environment. Am I missing a configuration
18:20 ganso step? I followed the Mirantis quickstart guide.
18:21 kupo24z You enable nova quotas in the fuel settings?
18:22 kupo24z xarses: looks like this image is unusable, I've gotten deployment timeouts every time. Controllers just time out when installing openstack
18:23 xarses can you reach the controller's from the admin network currently?
18:23 xarses or vise verse, can the controller reach the admin node?
18:24 xarses I was working with some one yesterday that was having problems because the network didn't like that the mac changed when the admin interface went from bare eth0 to the fw-admin bridge
18:24 kupo24z I can ping all 3 controllers br-fw-admin ip's from each controller
18:24 xarses =(
18:24 xarses at a loss then
18:24 xarses it should work
18:25 xarses and the fuel node?
18:25 e0ne joined #fuel
18:27 kupo24z Yup they can all ping the fuel nodes ip
18:28 kupo24z Odd though, I cant go from one node to another like I used to
18:28 ganso kupo24z: no Kupo I haven't set the Nova Quotes, the option is greyed out now, how can I enable it?
18:28 kupo24z root@node-10:~# ssh node-12 Warning: Permanently added 'node-12,192.168.0.5' (ECDSA) to the list of known hosts. Permission denied (publickey).
18:29 angdraug joined #fuel
18:29 kupo24z Im'a reset this and try without entering a syslog server and ssh key
18:32 kupo24z xarses: any way we could put the alias'ed node name in /etc/hosts puppet manifest?
18:36 e0ne joined #fuel
18:37 kupo24z ganso: Not sure outside of redployment
18:38 kupo24z what do you have for quota_driver in nova.conf?
18:42 kupo24z on a nova quotas=on env i have quota_driver=nova.quota.DbQuotaDriver but you might need other changes
18:49 kupo24z xarses: what format should the sshkey be in for the input field?
18:49 e0ne joined #fuel
18:49 kupo24z ie ssh-rsa $key $name?
18:50 xarses kupo24z: when fuel completes all of the node names should be in /etc/hosts for the cloud
18:50 xarses or do you mean on the fuel node its self?
18:50 kupo24z On the fuel settings page
18:51 kupo24z re: ssh key
18:51 xarses kupo24z: the same format as you put in authorized_keys
18:58 xarses kupo24z: who's /etc/hosts are you expecting to be populated
18:59 kupo24z fuel master & each node
18:59 kupo24z eg 192.168.0.6     node-24 node-24.$domain $alias
19:02 kupo24z where $alias is the label for each node entered in cluster/$num/nodes on the GUI
19:14 adanin joined #fuel
19:14 kupo24z Hmm, looks like its not hanging up on anything anymore after I deleted my syslog ip & ssh key
19:26 obcecado joined #fuel
19:26 obcecado hi
19:35 PMD_Kev joined #fuel
19:35 * PMD_Kev waves
19:36 PMD_Kev This may be an OpenStack-specific question -- but we're considering switching from a hosted Grizzly to a Fuel-based self-hosted Icehouse...
19:36 PMD_Kev Wondering if rsync-ing the virtuals is possible between the two.
19:37 PMD_Kev VMware could, in theory, import the exported .vdi from an OpenStack virtual.  I'm looking for the same type of migration path (as opposed to rebuilding everything)
19:39 adanin joined #fuel
19:52 kupo24z PMD_Kev: are the storage methods and virtualization types the same on both installs?
19:52 kupo24z PMD_Kev: standard method is to import the VM snapshot into glance on the new location
19:53 PMD_Kev Yes -- same on both.  LVM Cinder
19:53 PMD_Kev So I'd have to create a Glance image for each machine I wanted to preserve?  There are 30+
19:54 PMD_Kev I was hoping for something that was rsync imgold -> imagnew, then import it to the new compute to recognize it.
19:55 PMD_Kev I guess I can delete the glance when I'm done.  I'll check the googles and see if there's an alternative, but it's nice to know that it's possible.
19:56 PMD_Kev (or can I delete the glance image when I'm done?)
19:56 PMD_Kev I want to avoid having that show up in my image inventory for each server.
19:56 kupo24z As long as you dont have some kind of parent<->child relationship on your VM image creation you should be able to delete the source glance image
19:57 kupo24z however recovery mode would obviously fail
19:57 kupo24z Doing it manually can be possible but it involves editing the DB and a few more manual processes
19:58 PMD_Kev These images are all bootstrapped by Chef --it might be easier to just rebuild new images and drop the application to them.
19:58 PMD_Kev I'm really just trying to make a business case how we can save thousands with VMware by importing the old servers to the like environment.
19:58 PMD_Kev (save thousands compared to VMware, that is)
19:59 kupo24z https://blueprints.launchpad.net/cinder/+spec/add-export-import-volumes
19:59 kupo24z I imagine you could take that process and attempt it manually
20:04 kupo24z xarses: looks like its getting caught up on ceilometer
20:09 dilyin joined #fuel
20:11 meow-nofer_ joined #fuel
20:18 mattgriffin joined #fuel
20:22 HeOS joined #fuel
20:38 kupo24z made a bug about it https://bugs.launchpad.net/fuel/+bug/1360437
20:45 ByPasS joined #fuel
21:19 kupo24z Gunna try again with build 90
21:19 kupo24z looks like that might have some sql test fixes
21:21 e0ne joined #fuel
21:32 potato_farmer joined #fuel
21:32 potato_farmer Hi, is it safe to rename the hostnames set by fuel? (node1, node2, etc) ?
21:32 potato_farmer Does it refer to the hostnames for anything?
22:03 angdraug yes it does, so no, it's not safe
22:03 angdraug e.g. ceph.conf refers to host names
22:13 potato_farmer angdraug: thank you, one more question, does fuel actually use the public IPs on the nodes for anything? Or is all communication over the admin, management, storage, etc networks?
22:23 mattgriffin joined #fuel
22:46 angdraug potato_farmer: https://bugs.launchpad.net/bugs/1272349
23:22 nerdonmountain joined #fuel
23:56 mattgriffin joined #fuel

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