*** TheOsprey has quit IRC | 00:00 | |
*** lzyeval has joined #openstack | 00:02 | |
*** rnorwood has joined #openstack | 00:10 | |
*** JStoker has quit IRC | 00:11 | |
*** JStoker has joined #openstack | 00:14 | |
*** JStoker has quit IRC | 00:21 | |
Kiall | vyatta | 00:26 |
---|---|---|
*** JStoker has joined #openstack | 00:27 | |
*** n8 has joined #openstack | 00:32 | |
*** n8 is now known as Guest38394 | 00:32 | |
*** pradeep has joined #openstack | 00:32 | |
*** ahasenack has quit IRC | 00:33 | |
*** matiu2 has joined #openstack | 00:36 | |
*** rnorwood has quit IRC | 00:38 | |
*** blur has joined #openstack | 00:41 | |
*** JesperA has quit IRC | 00:44 | |
*** troya has joined #openstack | 00:44 | |
*** Guest38394 has quit IRC | 00:51 | |
*** RicardoSSP has joined #openstack | 01:01 | |
*** RicardoSSP has joined #openstack | 01:01 | |
coli | kiall: have you seen my comment about your ec2_dmz_host in nova.conf from your scripts ? | 01:02 |
Kiall | nope must have been lost in the netsplits! | 01:03 |
coli | you should set ec2_dmz_host to ip of a compute host and not to a ip of a controller, otherwise instance when contacting nova-api on the controler is not going to get the metadata (when compute and controler on different host) | 01:04 |
coli | you have nova-api on your computes but they are not used by instanaces | 01:04 |
Kiall | that doesnt sound right, since there could be 1000's of compute hosts? | 01:04 |
coli | and each compute host has nova-api which directly talks to mysql | 01:05 |
coli | just to show you what is happening: | 01:05 |
*** miclorb_ has quit IRC | 01:05 | |
coli | 1) instance when starting tries to contact 169.254.169.254 | 01:06 |
coli | 2) PREROUTING is DNAT'ing that request to controler ip and port 8773 | 01:06 |
coli | 3) routing takes place to go out of eth0 to controler host | 01:06 |
coli | 4) as this is remote host POSTROUTING takes place which is SNAT'ing the source to ip of compute host | 01:07 |
coli | 5) nova-api receives the request, and is looking up metadata for an instance based on the source ip address (which is not that of an instance but of an compute host) and doesn't find any | 01:08 |
*** blur has quit IRC | 01:08 | |
coli | if you set ec2_dmz_host to compues ip, then it contacts local nova-api and as it's local interface then no POSTROUTING takes place and source address is not nat'ed | 01:09 |
coli | ufff... | 01:09 |
*** woleium has quit IRC | 01:09 | |
Kiall | Ah yea.. I've actually got that in my 2nd+ node's nova.conf already .. Looks like its missing from the scripts! | 01:10 |
*** RicardoSSP has quit IRC | 01:11 | |
coli | overall your scripts are very nice and helped us a lot in undrestanding what is happening. thank you. | 01:11 |
Kiall | no problem :) | 01:12 |
Kiall | BTW - In your searching, you didnt find a way to avoid nova-api on each compute node did you? ;) | 01:12 |
* Kiall keeps typing into the wrong keyboard | 01:13 | |
*** n8 has joined #openstack | 01:13 | |
Kiall | (have a second PC at my desk for a few mins) | 01:13 |
*** n8 is now known as Guest51757 | 01:14 | |
coli | no in your solution, as you are using eth0 (public interface) to communicate with the controller, what causes all communication to be snat'ed | 01:14 |
Kiall | Humm, I wasnt aware any of the nova rules we're tied to a specific interface | 01:15 |
coli | if you had separate network just to communicate with conroler then you can do it using dmz_cidr parameter, what would cause to accept the packet before it's snated and controler is going to know where to send reply to the privae ip address of the instance | 01:15 |
coli | "tied to specific interface" ? I didn't get that | 01:16 |
Kiall | as in, none of the iptables rules have "-i <some interface>" | 01:16 |
coli | however wormman is running nova-api on each compute host as well as he has discovered performance issues when using just one nova-api | 01:17 |
coli | kiall: none of them have -i and we are not tied | 01:17 |
Kiall | WormMan: noticed the issue with launching 200 instances at the same time :) | 01:17 |
Kiall | s/:// | 01:17 |
coli | kiall we hope to get there... don't we ? ;-) | 01:18 |
Kiall | Just digging through the code, and yea.. I'm setting dmz_cidr way off | 01:19 |
coli | let me thnik. | 01:19 |
coli | no ideas right now how to do it otherwise in your setup | 01:20 |
cloudgeek | going for exam | 01:21 |
coli | in my opinnion the main architectural/desing flaw is that nova-api searches for metadata based on source ip address. or there is something what i'm missing | 01:21 |
cloudgeek | study whole night bye openstack | 01:21 |
Kiall | yea, eth1 should get an IP in some dmz_cidr, and that IP should be used in the ec2_dmz_host setting .. (Right now, I force my eth1's up without an IP) | 01:21 |
coli | cloudgeek goodluck | 01:21 |
Kiall | good luck | 01:21 |
cloudgeek | coli:Kiall:thanks | 01:21 |
coli | kiall you don't need that, as nova-api listens on all interfaces you can set ec2_dmz_host to public ip of the compute host | 01:22 |
*** livemoon has joined #openstack | 01:23 | |
Kiall | sure, and set dmz_cidr to the controllers public subnet, but my floating IPs are in the same subnet.. so setting dmz_cidr to the public subnet will end badly | 01:23 |
coli | you don't need to set dmz_cidr in my opinnion as the ec2_dmz_host is local so no POSTROUTING is happening and no source address rewriting (snat) | 01:24 |
coli | all calls are going to be to local nova-api which in trun will talk to mysql directly | 01:24 |
Kiall | Ah, I'm talking about when using 1 nova-api instance (or, dedicated nova-api servers) rather than co-locating it on every compute node | 01:24 |
coli | then you need to do it | 01:25 |
coli | but remember that packets going to controler will have source address of an instance (its fixed ip) so your controler needs to know how to route those | 01:25 |
Kiall | 1 nova-api per nova-compute is not really necessary IMO, even for larger deployments.. The added connections overhead to MySQL would be a PITA vs a pool of nova-api servers... | 01:26 |
coli | your setup is not designed for large deplyments as it can only support up to approx 4080 tenants | 01:26 |
coli | you are using vlan network mode | 01:26 |
Kiall | Yea - thats a hard limit alright ;) | 01:27 |
coli | i'm just working out how to use flatDHCP instead in order to get rid of vlan's and have one large network on the fixed_ip side. | 01:27 |
*** pradeep has quit IRC | 01:28 | |
Kiall | coli: feel free to send a pull request with any changes to the scripts, I'll try and merge them in as an option... People keep asking about flat DHCP ;) | 01:28 |
coli | I have no idea what a pull request is :-) it's my first time I'm using git :-( | 01:29 |
Kiall | Github has good git guides ;) | 01:29 |
coli | when I get somewhere reasonably working I will let you know and send you my findings :-) | 01:29 |
Kiall | Anyway - I've like another 200 firewall rules to check before heading home :( | 01:29 |
coli | get focused then | 01:30 |
troya | hi all | 01:30 |
coli | hi | 01:31 |
troya | coli, do you have more experience use Flatdhcp ? | 01:33 |
coli | troya: not yet, just starting | 01:33 |
troya | i have some problem with my instances now, i cannot ping and accessing ssh into.whereas i have configure security rule to allowing ssh and icmp | 01:35 |
coli | did you configure the policies through dashboard ? | 01:36 |
*** woleium has joined #openstack | 01:36 | |
coli | did you assign floating ip to the instance ? | 01:36 |
Kiall | set interface openvpn vtun3 mode site-to-site | 01:36 |
Kiall | dooh | 01:36 |
troya | i don't install dashboard | 01:38 |
coli | ok | 01:38 |
coli | did you have a look at iptables and its counters ? | 01:38 |
troya | and i haven't configure float ip to isntances | 01:38 |
troya | i think i haven't configure any rules on ip tables | 01:39 |
*** winston-d has joined #openstack | 01:41 | |
troya | any aide ? | 01:41 |
coli | troya: nova-network does the configuring of iptables for you | 01:43 |
coli | check if it did its job properly | 01:44 |
troya | this is nova-network.log , it's like normal | 01:47 |
*** krow has quit IRC | 01:50 | |
*** livemoon has left #openstack | 01:50 | |
*** rnorwood has joined #openstack | 01:51 | |
*** debo-os has joined #openstack | 01:56 | |
troya | this is pastebin of nova-compute.log >> http://pastebin.com/YHzwSwTW | 01:58 |
*** crescendo has joined #openstack | 02:02 | |
*** crescendo has joined #openstack | 02:02 | |
*** lzyeval_ has joined #openstack | 02:04 | |
*** lzyeval has quit IRC | 02:04 | |
*** rnorwood has joined #openstack | 02:09 | |
troya | Kiall, can you explains me, if i install nova without create volume ? | 02:09 |
*** osier has joined #openstack | 02:10 | |
*** songyan has joined #openstack | 02:11 | |
*** rustam has quit IRC | 02:13 | |
*** songyan has quit IRC | 02:14 | |
*** _diana_1 has left #openstack | 02:21 | |
*** lzyeval_ has quit IRC | 02:30 | |
*** lzyeval has joined #openstack | 02:30 | |
*** shang has quit IRC | 02:30 | |
*** dolphm has joined #openstack | 02:32 | |
*** dolphm has quit IRC | 02:37 | |
*** MarkAtwood has quit IRC | 02:43 | |
*** MarkAtwood has joined #openstack | 02:44 | |
*** MarkAtwood has left #openstack | 02:44 | |
*** MarkAtwood has joined #openstack | 02:45 | |
*** MarkAtwood has left #openstack | 02:46 | |
*** rods has quit IRC | 02:47 | |
*** shang has joined #openstack | 02:48 | |
*** CaptTofu has joined #openstack | 02:52 | |
*** debo-os has quit IRC | 02:53 | |
*** songyan has joined #openstack | 03:00 | |
*** deshantm_laptop has joined #openstack | 03:02 | |
*** lzyeval has quit IRC | 03:11 | |
*** lzyeval has joined #openstack | 03:11 | |
*** rnorwood has quit IRC | 03:16 | |
*** pradeep1 has joined #openstack | 03:20 | |
*** cloudgeek has quit IRC | 03:24 | |
*** hadrian has quit IRC | 03:24 | |
*** rnorwood has joined #openstack | 03:36 | |
*** miclorb_ has joined #openstack | 03:37 | |
*** CaptTofu has quit IRC | 03:42 | |
lzyeval | I've just finished reading most of the important code of nova. | 03:44 |
lzyeval | Anybody interested in talking about it, email me. | 03:44 |
lzyeval | lzyeval gmail | 03:44 |
*** koolhead17 has quit IRC | 03:53 | |
*** sdake has quit IRC | 03:53 | |
*** map_nw_ has joined #openstack | 03:55 | |
*** rnorwood has quit IRC | 03:57 | |
*** map_nw has quit IRC | 03:58 | |
*** deshantm_laptop_ has joined #openstack | 04:02 | |
*** rnorwood has joined #openstack | 04:02 | |
*** deshantm_laptop has quit IRC | 04:05 | |
*** sdake has joined #openstack | 04:05 | |
*** gohko_air has quit IRC | 04:13 | |
*** deshantm_laptop_ has quit IRC | 04:21 | |
*** deshantm_laptop has joined #openstack | 04:21 | |
*** Arminder has quit IRC | 04:35 | |
*** Arminder has joined #openstack | 04:36 | |
*** supriya has joined #openstack | 04:44 | |
*** rsampaio has quit IRC | 04:48 | |
*** wariola has quit IRC | 04:57 | |
*** dotdevops has joined #openstack | 05:02 | |
*** rnorwood has quit IRC | 05:08 | |
*** wariola has joined #openstack | 05:16 | |
*** dysinger has joined #openstack | 05:18 | |
*** wariola has quit IRC | 05:19 | |
*** wariola has joined #openstack | 05:20 | |
*** dysinger has quit IRC | 05:20 | |
*** dysinger has joined #openstack | 05:20 | |
*** Guest51757 has quit IRC | 05:21 | |
*** n8 has joined #openstack | 05:22 | |
*** n8 is now known as Guest35322 | 05:22 | |
*** dysinger1 has joined #openstack | 05:24 | |
*** mjfork has quit IRC | 05:25 | |
*** wariola has quit IRC | 05:25 | |
*** dysinger has quit IRC | 05:27 | |
*** dysinger1 has quit IRC | 05:31 | |
*** pradeep1 has quit IRC | 05:32 | |
*** dysinger has joined #openstack | 05:33 | |
*** wariola has joined #openstack | 05:38 | |
*** Guest35322 has quit IRC | 05:40 | |
*** shaon has joined #openstack | 05:46 | |
*** dysinger has quit IRC | 05:50 | |
*** localhost has quit IRC | 06:00 | |
*** localhost has joined #openstack | 06:01 | |
*** vernhart has joined #openstack | 06:18 | |
*** lzyeval has quit IRC | 06:18 | |
*** arBmind has joined #openstack | 06:28 | |
*** MarkAtwood has joined #openstack | 06:32 | |
*** pradeep has joined #openstack | 06:34 | |
*** dnjaramba has quit IRC | 06:38 | |
HugoKuo_ | is there have API for checking the security_group of an instance? | 06:42 |
*** redconnection has quit IRC | 06:45 | |
*** shaon has quit IRC | 06:46 | |
*** koolhead17 has joined #openstack | 06:49 | |
*** redconnection has joined #openstack | 06:50 | |
*** miclorb_ has quit IRC | 06:53 | |
*** sdake has quit IRC | 06:53 | |
*** debo-os has joined #openstack | 06:54 | |
*** wariola has quit IRC | 06:56 | |
*** tjikkun has quit IRC | 06:57 | |
*** tjikkun has joined #openstack | 06:58 | |
*** redconnection has quit IRC | 06:59 | |
*** kaigan_ has joined #openstack | 06:59 | |
*** mindpixel has joined #openstack | 07:00 | |
*** redconnection has joined #openstack | 07:02 | |
*** tjikkun has quit IRC | 07:03 | |
*** TheOsprey has joined #openstack | 07:05 | |
*** tjikkun has joined #openstack | 07:05 | |
*** redconnection has quit IRC | 07:07 | |
*** wariola has joined #openstack | 07:09 | |
*** nerens has joined #openstack | 07:10 | |
*** janpy has quit IRC | 07:15 | |
*** negronjl has quit IRC | 07:19 | |
*** livemoon has joined #openstack | 07:23 | |
*** guigui has joined #openstack | 07:26 | |
*** negronjl has joined #openstack | 07:32 | |
*** yamahata_ has quit IRC | 07:33 | |
*** tsuzuki_ has joined #openstack | 07:34 | |
*** janpy has joined #openstack | 07:34 | |
*** deshantm_laptop has quit IRC | 07:36 | |
*** ritesh has joined #openstack | 07:37 | |
ritesh | my nova.compute gives error:-Unable to load the virtualization driver: No module named XenAPI | 07:38 |
ritesh | even i had defined --connection_type=xenapi --xenapi_connection_url=http://169.144.104.73 --xenapi_connection_username=root --xenapi_connection_password=red32hat | 07:39 |
*** redconnection has joined #openstack | 07:39 | |
*** yamahata has joined #openstack | 07:41 | |
*** debo-os has quit IRC | 07:43 | |
*** thickskin has joined #openstack | 07:45 | |
thickskin | is there someone? | 07:46 |
*** pradeep has quit IRC | 07:53 | |
*** pradeep has joined #openstack | 07:53 | |
*** pradeep has quit IRC | 07:56 | |
*** pradeep has joined #openstack | 07:56 | |
*** wariola has quit IRC | 08:00 | |
*** cw has quit IRC | 08:01 | |
*** debo-os has joined #openstack | 08:02 | |
*** pradeep has quit IRC | 08:02 | |
*** cw has joined #openstack | 08:06 | |
*** arBmind has quit IRC | 08:08 | |
*** shang has quit IRC | 08:12 | |
*** nati2_ has joined #openstack | 08:16 | |
*** reidrac has joined #openstack | 08:16 | |
*** nati2 has quit IRC | 08:18 | |
*** dobber has joined #openstack | 08:19 | |
*** dev_sa has joined #openstack | 08:19 | |
*** Vek has quit IRC | 08:22 | |
*** nerens has quit IRC | 08:25 | |
*** shang has joined #openstack | 08:31 | |
*** Russell has joined #openstack | 08:31 | |
*** Russell is now known as Guest88 | 08:32 | |
zykes- | notmyname: around ? | 08:36 |
*** vdo has joined #openstack | 08:37 | |
*** woleium has quit IRC | 08:38 | |
*** livemoon has left #openstack | 08:42 | |
zykes- | doh, probably not around for another few hours ;p | 08:47 |
*** miclorb_ has joined #openstack | 08:56 | |
*** oonersch has joined #openstack | 08:56 | |
zykes- | anyone here got clues on replication on swift? | 08:57 |
reidrac | what kind of clues? :) | 08:58 |
*** uksysadmin has joined #openstack | 08:59 | |
*** koolhead17 has quit IRC | 09:02 | |
*** janpy has quit IRC | 09:02 | |
*** dirkx has joined #openstack | 09:03 | |
*** thickskin has left #openstack | 09:03 | |
*** guigui has quit IRC | 09:05 | |
*** javiF has joined #openstack | 09:09 | |
*** tudor has joined #openstack | 09:11 | |
zykes- | reidrac: we're wanting to do | 09:15 |
zykes- | 2 nodes pr location | 09:15 |
zykes- | with 4 zones | 09:15 |
zykes- | does that sound reasonable ? | 09:15 |
reidrac | 2 nodes per location? | 09:15 |
zykes- | 2 servers | 09:15 |
zykes- | for replication | 09:16 |
zykes- | 2 in each DC | 09:16 |
zykes- | bad idea or ? | 09:16 |
reidrac | are you building a cluster split in several DC? | 09:16 |
reidrac | if you have a very good link between them, it shouldn't be a problem | 09:16 |
zykes- | do you need 4 nodes in a seperate cluster ? | 09:16 |
*** mu574n9 has quit IRC | 09:17 | |
reidrac | the zone is supposed to be the biggest point of failure you're going to deal with | 09:17 |
zykes- | so you can have like 2 zones in one dc and 2 in another ? | 09:18 |
reidrac | ie. a zone can be a server, each server with different power bar in a different rack | 09:18 |
reidrac | if you do that, and you lose 1 dc, you're out with 2 zones missing | 09:18 |
zykes- | so basically screwed ? | 09:18 |
reidrac | the zones are used for replication | 09:19 |
zykes- | yeah | 09:19 |
zykes- | but like | 09:19 |
reidrac | you're guaranteed to have data replicated into different zones | 09:19 |
zykes- | we have 1 "data" room here in the office building and also a colocation datacenter | 09:19 |
zykes- | we are looking at having 2 servers (2 zones) pr location | 09:19 |
zykes- | would that be doable ? | 09:20 |
reidrac | as long as you have good connection between locations (bandwidth is vital!), yep | 09:20 |
zykes- | what happens if i loose the connection in between then for say 3-5 hours ? | 09:20 |
reidrac | you loose 2 zones in the cluster | 09:20 |
zykes- | but is the "replicas" still fine ? | 09:21 |
zykes- | or will i suffer from data loss | 09:21 |
reidrac | if you're using 4 zones with 3 replicas and lose 2 zones, you can't operate | 09:21 |
zykes- | :/ | 09:21 |
zykes- | so no swift for uss then | 09:21 |
reidrac | I wouldn't split the zones in different DCs at first | 09:22 |
tudor | anybody have experience setting up live migration? i set it up as per documentation and get a "TRACE: AMQPConnectionException: (503, u"COMMAND_INVALID - second 'channel.open' seen", (20, 10), 'Channel.open') | 09:22 |
reidrac | it's an expensive design choice and there's not big benefit IMHO | 09:22 |
*** wi3dzma has joined #openstack | 09:23 | |
zykes- | reidrac: wouldn't you need that for off site distasters ? | 09:23 |
reidrac | I think that if you lose a DC, something that SHOULD be very unlikely, it's fine your storage is unavailable ;) | 09:23 |
reidrac | zone = server it's usually a good idea, because hw can fail, and losing 1 server it's not impossible | 09:24 |
reidrac | but using different DCs as zones, that's a completely different level :D | 09:25 |
zykes- | reidrac: that's one thing, but what about having other copies of the data ? | 09:25 |
zykes- | i mean off site from the datacenter | 09:25 |
reidrac | you can have backups offsite | 09:25 |
zykes- | using swift or ? | 09:26 |
reidrac | as a recovery plan | 09:26 |
*** nacx has joined #openstack | 09:26 | |
reidrac | not using swift | 09:26 |
zykes- | that's kind of the deal, we're only within 8 km from the DC | 09:26 |
zykes- | so it's simple enough except from bw to replicate over the link | 09:26 |
reidrac | what are we talking about? | 09:26 |
reidrac | what kind of disaster? | 09:27 |
zykes- | reidrac: say a fire within the dc ? | 09:27 |
zykes- | has happened before :) | 09:27 |
reidrac | fire on the DC or it's just a matter of availability | 09:27 |
reidrac | ah, I see | 09:27 |
reidrac | I would recommend looking for a different DC | 09:27 |
reidrac | hehe | 09:27 |
zykes- | reidrac: not like that, but it's not impossible for it to happen | 09:27 |
*** matiu2 has quit IRC | 09:27 | |
*** guigui has joined #openstack | 09:29 | |
*** Arminder has quit IRC | 09:32 | |
zykes- | reidrac: can't one do cross dc replication then ? | 09:35 |
reidrac | you need lots of bandwidth between the storage nodes | 09:36 |
reidrac | can be done, probably not worth it | 09:36 |
*** pixelbeat has joined #openstack | 09:36 | |
*** katkee has joined #openstack | 09:36 | |
zykes- | so then basically it's useless | 09:36 |
zykes- | for what we wanted it for ;p | 09:36 |
zykes- | then my idea died, so sad :p | 09:38 |
*** dysinger has joined #openstack | 09:38 | |
reidrac | ask for a second opinion, I may be wrong | 09:38 |
reidrac | you know... on the internet nobody knows you're a dog ;) | 09:39 |
zykes- | :p | 09:39 |
uvirtbot | New bug: #894683 in nova "nova.service does not handle attribute specific exceptions and client hangs " [Undecided,New] https://launchpad.net/bugs/894683 | 09:41 |
uksysadmin | reidrac, I've just been reading this conversation - many people to site to site storage replication (we shift Tbs a day across the equiv of dark fiber) - what is large storage and lots of bandwidth? | 09:41 |
zykes- | uksysadmin: how many zones pr location ? | 09:42 |
uksysadmin | Do you know of any benchmarks/recommendations on this? | 09:42 |
zykes- | or so | 09:42 |
uksysadmin | we're not using Swift - was wondering what the impact of Swift would be | 09:42 |
reidrac | uksysadmin: it's just our experience | 09:42 |
*** nati2_ has quit IRC | 09:42 | |
Daviey | uksysadmin: what do you use for storage? | 09:43 |
uksysadmin | NetApp | 09:43 |
*** nati2 has joined #openstack | 09:43 | |
zykes- | ah ok | 09:43 |
*** Guest88 has quit IRC | 09:43 | |
uksysadmin | it was just out of curiosity - we've no plans to implement Swift but handy to know the design requirements and limits should we want to | 09:44 |
*** darraghb has joined #openstack | 09:47 | |
*** songyan has quit IRC | 09:47 | |
*** dev_sa has quit IRC | 09:49 | |
*** jedi4ever has joined #openstack | 09:49 | |
*** cloudgeek has joined #openstack | 09:53 | |
*** dev_sa has joined #openstack | 09:55 | |
*** koolhead11 has joined #openstack | 09:55 | |
*** debo-os has quit IRC | 09:56 | |
*** rustam has joined #openstack | 09:58 | |
*** debo-os has joined #openstack | 09:59 | |
*** lzyeval has joined #openstack | 10:01 | |
*** debo-os_ has joined #openstack | 10:02 | |
*** debo-os has quit IRC | 10:03 | |
*** nati2_ has joined #openstack | 10:04 | |
*** debo-os has joined #openstack | 10:04 | |
*** guynaor has joined #openstack | 10:06 | |
*** debo-os_ has quit IRC | 10:06 | |
*** nati2 has quit IRC | 10:06 | |
*** guynaor has left #openstack | 10:07 | |
*** cloudgeek has left #openstack | 10:13 | |
*** supriya has quit IRC | 10:17 | |
*** Vek has joined #openstack | 10:18 | |
*** guigui has quit IRC | 10:23 | |
*** dysinger has quit IRC | 10:23 | |
*** lzyeval has quit IRC | 10:24 | |
*** guigui has joined #openstack | 10:27 | |
*** ritesh has quit IRC | 10:27 | |
*** nati2_ has quit IRC | 10:32 | |
*** nati2 has joined #openstack | 10:33 | |
*** PotHix has joined #openstack | 10:33 | |
*** massoud has joined #openstack | 10:39 | |
*** supriya has joined #openstack | 10:42 | |
*** dev_sa has quit IRC | 10:46 | |
*** dev_sa has joined #openstack | 10:46 | |
zykes- | notmyname: you around now ? | 10:48 |
*** dirkx_ has joined #openstack | 10:50 | |
*** dirkx has quit IRC | 10:52 | |
*** dirkx_ is now known as dirkx | 10:52 | |
*** troya has quit IRC | 11:00 | |
*** binbash_ has quit IRC | 11:01 | |
*** mgoldmann has joined #openstack | 11:01 | |
*** binbash_ has joined #openstack | 11:01 | |
*** osier has quit IRC | 11:07 | |
*** lzyeval has joined #openstack | 11:12 | |
*** nati2_ has joined #openstack | 11:16 | |
*** nati2 has quit IRC | 11:18 | |
*** matiu_ has joined #openstack | 11:23 | |
*** matiu_ has joined #openstack | 11:23 | |
*** yshh has joined #openstack | 11:31 | |
*** koolhead11 has left #openstack | 11:33 | |
*** GheRivero has quit IRC | 11:36 | |
*** dirkx has quit IRC | 11:41 | |
*** miclorb_ has quit IRC | 11:42 | |
*** GheRivero has joined #openstack | 11:42 | |
*** lionel has quit IRC | 11:44 | |
*** lionel has joined #openstack | 11:45 | |
*** rsampaio has joined #openstack | 12:01 | |
*** matiu_ has quit IRC | 12:01 | |
*** lzyeval has quit IRC | 12:06 | |
*** lzyeval has joined #openstack | 12:07 | |
*** Arminder has joined #openstack | 12:12 | |
*** javiF has quit IRC | 12:13 | |
*** dirkx has joined #openstack | 12:15 | |
*** lzyeval has quit IRC | 12:22 | |
zykes- | notmyname: ? ;p | 12:23 |
*** supriya has quit IRC | 12:24 | |
*** mnabil_ has joined #openstack | 12:24 | |
*** praefect_ has quit IRC | 12:30 | |
*** praefect has joined #openstack | 12:30 | |
*** dolphm has joined #openstack | 12:34 | |
*** marcuz has joined #openstack | 12:38 | |
*** dolphm has quit IRC | 12:38 | |
*** rsampaio has quit IRC | 12:43 | |
*** Ramonster has joined #openstack | 12:48 | |
*** guigui has quit IRC | 12:55 | |
*** tsuzuki_ has quit IRC | 12:56 | |
uvirtbot | New bug: #894733 in openstack-ci "Jenkins periodic jobs no longer run" [Critical,Confirmed] https://launchpad.net/bugs/894733 | 12:56 |
*** dysinger has joined #openstack | 12:57 | |
*** mgoldmann has quit IRC | 12:59 | |
*** rustam has quit IRC | 13:04 | |
*** JStoker has quit IRC | 13:05 | |
*** troya has joined #openstack | 13:06 | |
uvirtbot | New bug: #894741 in glance "'glance index' and 'glance details' empty" [Undecided,New] https://launchpad.net/bugs/894741 | 13:11 |
*** supriya has joined #openstack | 13:15 | |
*** JStoker has joined #openstack | 13:16 | |
*** andersem has joined #openstack | 13:20 | |
andersem | hi, when trying to save an object via object.load_from_filename() in python-cloudfiles I sometimes get a 400-error like this: ResponseError: 400: Bad request syntax ('\xaa\xaa\x00\x00\x014\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaaE\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa | 13:22 |
andersem | oops, a little long | 13:23 |
andersem | anyway, when retrying to upload the file, it usually succeeds | 13:23 |
andersem | it's my own swift-install I'm uploading to | 13:24 |
andersem | and it has worked perfectly for a long time | 13:24 |
andersem | anyone got an idea where I could find the reason for this error? | 13:25 |
andersem | the syslog doesn't give me a lot of help, except for listing that a 400 and a 404 was sent | 13:26 |
*** foexle has joined #openstack | 13:26 | |
*** mnabil_ has quit IRC | 13:26 | |
foexle | hiho | 13:26 |
*** mjfork has joined #openstack | 13:28 | |
*** andersem has quit IRC | 13:39 | |
*** omsynge has joined #openstack | 13:39 | |
*** rods has joined #openstack | 13:40 | |
omsynge | hello everyone, I keep getting stuck on the first stage of installing an open stack deve server, nova-manage network create novanetwork 192.168.87.200/24 1 64 | 13:41 |
omsynge | Command failed, please check log for more info | 13:41 |
omsynge | could anyone give me a hand on the early stages of networking setup for open stack | 13:41 |
asksol | Does OpenStack target Python 2.4? | 13:42 |
omsynge | 2011-11-25 14:41:09,818 CRITICAL nova [-] --bridge is required to create a network. | 13:42 |
omsynge | (nova): TRACE: Traceback (most recent call last): | 13:42 |
omsynge | (nova): TRACE: File "/usr/bin/nova-manage", line 2340, in <module> | 13:42 |
omsynge | (nova): TRACE: main() | 13:42 |
omsynge | (nova): TRACE: File "/usr/bin/nova-manage", line 2328, in main | 13:42 |
omsynge | (nova): TRACE: fn(*fn_args, **fn_kwargs) | 13:42 |
omsynge | (nova): TRACE: File "/usr/bin/nova-manage", line 768, in create | 13:42 |
omsynge | (nova): TRACE: raise exception.NetworkNotCreated(req='--bridge') | 13:42 |
omsynge | (nova): TRACE: NetworkNotCreated: --bridge is required to create a network. | 13:42 |
omsynge | root@sky403:~# python | 13:42 |
omsynge | Python 2.7.2+ (default, Oct 4 2011, 20:06:09) | 13:42 |
*** supriya has quit IRC | 13:42 | |
omsynge | root@sky403:~# lsb_release -a | 13:42 |
omsynge | No LSB modules are available. | 13:42 |
omsynge | Distributor ID: Ubuntu | 13:42 |
omsynge | Description: Ubuntu precise (development branch) | 13:42 |
omsynge | Release: 12.04 | 13:42 |
omsynge | Codename: precise | 13:42 |
omsynge | or I can be RHEL 6.0 | 13:42 |
omsynge | which ever OS you like | 13:43 |
foexle | check you nova options, if the bridge flag is set | 13:43 |
omsynge | # cat /etc/nova/nova.conf | 13:44 |
omsynge | --dhcpbridge_flagfile=/etc/nova/nova.conf | 13:44 |
omsynge | --dhcpbridge=/usr/bin/nova-dhcpbridge | 13:44 |
omsynge | --network_manager=nova.network.manager.FlatDHCPManager | 13:44 |
omsynge | --flat_network_dhcp_start=192.168.87.200 | 13:44 |
omsynge | --public_interface=eth0 | 13:44 |
omsynge | --logdir=/var/log/nova | 13:44 |
omsynge | --state_path=/var/lib/nova | 13:44 |
omsynge | --lock_path=/var/lock/nova | 13:44 |
omsynge | --force_dhcp_release | 13:44 |
omsynge | --use_deprecated_auth | 13:44 |
omsynge | --iscsi_helper=tgtadm | 13:44 |
omsynge | --verbose | 13:44 |
*** guigui1 has joined #openstack | 13:44 | |
omsynge | sould I add soem thign like --bridge=eth0 | 13:45 |
omsynge | oh tried that on the command line and it seems to work | 13:46 |
omsynge | nova-manage network create novanetwork 192.168.87.200/24 1 64 --bridge=eth0 | 13:46 |
omsynge | seems to work | 13:46 |
omsynge | (having been playing with libvirt and simple scripts for soem time I am expecting the cloud to be a networking pain in ***) | 13:47 |
*** dev_sa has left #openstack | 13:49 | |
tudor | anybody has experience with setting up live migration on diablo? | 13:56 |
*** GheRivero has quit IRC | 13:58 | |
omsynge | sorry tudor not that far, and assume you need a shared file store fro migration, if you do I think its not applicable to the size I want to scale our services too, but understand for smaller sites this may be practical | 13:59 |
omsynge | We have a XEN based cloud doing this and netapp back ends, and performance is nto good enough on IO, so we went fibre channel and still performance is not great | 14:00 |
tudor | thx omsynge, i just ran into some nfs permission problems i don't see getting arround, i was wondering if anybody did | 14:00 |
omsynge | Hi Tudor, We never tried NFS for long, but went , iSCSI -> fibrechannel | 14:05 |
omsynge | (its one of nice things about working in Public Science, we are allowed to play with big hardware and to talk about things that went wrong and not hide them) | 14:08 |
coli | --flat_network_dhcp_start i ddepriciated afaik | 14:09 |
omsynge | oh thanks coli | 14:09 |
omsynge | I have got further than ever before in open stack now, am begging to fight the beast, seems not as bad as early days of the grid projects, but still not got it all running nice yet | 14:13 |
omsynge | thanks | 14:14 |
notmyname | zykes-: what time zone are you in? | 14:14 |
*** cloudgeek has joined #openstack | 14:14 | |
cloudgeek | Hi all | 14:14 |
tudor | omsynge: so you got live migration to work with nfs on diablo release? what about iscsi, what distribuited file system did you use? | 14:21 |
*** mcclurmc has quit IRC | 14:22 | |
*** mcclurmc has joined #openstack | 14:23 | |
cloudgeek | looking for a student project in cloud computing (openstack+hadoop)is there any ? | 14:24 |
Glacee | you mean runnning hadoop clusters in a nova setup? | 14:26 |
*** rsampaio has joined #openstack | 14:31 | |
omsynge | sorry tudor no we got live migration working with the citrix xen system working | 14:33 |
omsynge | This said, I dont think NFS is a very scalable solution to back end VM storage until NFS 4.1 is standard and only then it may change | 14:35 |
*** deshantm_laptop has joined #openstack | 14:35 | |
*** dolphm has joined #openstack | 14:35 | |
*** hggdh has quit IRC | 14:38 | |
zykes- | notmyname: gmt+1 ;) | 14:39 |
*** rustam has joined #openstack | 14:41 | |
*** hggdh has joined #openstack | 14:42 | |
uvirtbot | New bug: #893961 in nova "nova-api does not use flags of osapi_host " [Undecided,Invalid] https://launchpad.net/bugs/893961 | 14:42 |
coli | I have a network of 16 fixed_ip's assigned to tenant (project), I did create and terminate number of instances, now I have no instances running, however when creating a new one I get NoMoreFixedIps exception. | 14:43 |
coli | looked into source code and even worked out which sql query checks for the fixed ips in the mysql, however cannot work out why I get this exception. any ideas ? | 14:43 |
cloudgeek | Glacee:I used Hadoop ?! but not with openstack but i will try to experiment after my exams | 14:44 |
*** pradeep has joined #openstack | 14:44 | |
cloudgeek | coli:project ! | 14:44 |
cloudgeek | where is vidd | 14:45 |
coli | coudgeek: what do you mean ? | 14:45 |
uvirtbot | New bug: #894776 in horizon "run_test.sh does not work without virtualenv" [Undecided,New] https://launchpad.net/bugs/894776 | 14:46 |
cloudgeek | coli: is project avail ? | 14:46 |
coli | couldgeek: still don't undrestand your question, sorry :-) | 14:47 |
*** _rfz has joined #openstack | 14:47 | |
praefect | coli: --update_dhcp_on_disassociate --force_dhcp_release --fixed_ip_diassociate_timeout | 14:47 |
praefect | coli: these and a patch of manager.py did the trick for me | 14:48 |
*** helfrez has joined #openstack | 14:48 | |
cloudgeek | coli: i just asked about a project ! then you post something about project ! that is misconception my be now it's okay | 14:48 |
*** TheOsprey has quit IRC | 14:48 | |
praefect | coli: do you use auto_assign ? | 14:48 |
praefect | of floating | 14:48 |
coli | praefect: not atm | 14:48 |
coli | praefect: which patch did you use ? | 14:48 |
_rfz | praefect, how do you configure auto_assign of floating addresses? | 14:49 |
praefect | coli: I'll look it up, it is a single line to change in network/manager.py but it makes a lot a difference when you use auto assign of floating | 14:49 |
cloudgeek | what is meaning of round shape blue point infront of some guy in IRC like mtaylor , redbo soren ! | 14:49 |
cloudgeek | can anybody tell about what does that mean are they primes in irc ! | 14:49 |
praefect | --auto_assign_floating_ip=true but if you don't patch manager.py your floating will never release, with the patch it's flawless! | 14:49 |
coli | prafect: my problem is with fixed_ips at the moment | 14:50 |
_rfz | praefect, where can I get the patch? | 14:50 |
coli | praefect: but if you can find the one line patch I will be grateful | 14:50 |
praefect | coli: I'm looking for it, wait a sec I'll just do a diff it'll pop right back | 14:51 |
_rfz | coli, what problem you having with fixed_ip? | 14:51 |
coli | _rfz: I have a network of 16 fixed_ip's assigned to tenant (project), I did create and terminate number of instances, now I have no instances running, however when creating a new one I get NoMoreFixedIps exception. | 14:51 |
_rfz | have you got --force_dhcp_release=True on all your nova.conf? | 14:53 |
_rfz | coli, are you using kiall's repo? | 14:54 |
praefect | coli: https://review.openstack.org/#patch,sidebyside,1328,3,nova/network/manager.py this is what I changed, I'm using rpm packages for nova so our code may be different, for example of the 2 changes listed in this URL, one was already in place in my code, not the other.. I applied this and auto_assign works flawlessly | 14:55 |
coli | _rfz: it's not a repo, but started with his scripts. | 14:55 |
*** sdake has joined #openstack | 14:55 | |
_rfz | so how did you install nova-compute? | 14:55 |
coli | _rfz: corrected his ec2_dmz_host setting as well | 14:56 |
coli | ./nova-compute.sh | 14:56 |
praefect | _rfz: since I'm on centos, I don't have access to the dhcp_relase binary which you guys get access to in ubuntu, hence I just generate errors currently when trying to force dhcp release of fixed ip, but to answer your question, this error only appears on my network node so you just need this switch on your controller(network) I guess | 14:56 |
coli | and that is all | 14:56 |
coli | remember to change ec2_dmz_host to a public ip of your compute host | 14:56 |
praefect | all works well without dhcp_release, the force-dhcp-release is just a way to speed things up, upon termination the fixed ip will be release automatically | 14:57 |
praefect | if not there is a timeout and the fixed ip will be cleaned eventually | 14:57 |
_rfz | praefect, what guide did you use to install openstack on centos? | 14:57 |
coli | praefect: do you have nova-network on a controler or on computes ? | 14:57 |
_rfz | you need them on both | 14:57 |
praefect | _rfz: ALL OF THEM! there is not much but start with the official one (search RHEL on openstack.org) | 14:58 |
_rfz | praefect, hehe | 14:58 |
praefect | coli: I have nova-network only on my controller, (I'm not using multi_host) | 14:58 |
troya | hi all | 14:59 |
coli | ok | 14:59 |
praefect | coli: I think you need 2 NICs on all compute nodes if you want them to be network controllers (and use floating ips) | 14:59 |
praefect | which I don't | 14:59 |
coli | praefect: I have three hosts are present, one controler and two computes. each has two nics | 15:00 |
praefect | I channel everything to my controller (running nova-network) and traffix goes out through its public NIC (floating IP) | 15:00 |
coli | anybody attending DCD next week ? | 15:00 |
praefect | coli: then in this setup you could run multi_host=T (when creating your network) this will spawn a nova-network on all nodes and route traffic individually, the goal is that with this setup, if you loose the controller, spawned instances will still have access to the network (internet) | 15:01 |
coli | praefect: that is how I got it setup right now. | 15:01 |
praefect | while in my case, if I loose the controller, even if the instances are up they are effectively cut off from the network | 15:01 |
*** CaptTofu has joined #openstack | 15:02 | |
andyandy_ | do you know if I can have more then one nova-volume ? thanks. | 15:03 |
praefect | coli: I have 2 clouds here, one with 2 powerful servers (24 cores each), and one cloud with desktops (8 cores xeon)... I'm going to play with zones and scheduling today)... but anyway, my setup is really nice now, no more IP management problems, not more lock file problems (only affecting centOS) | 15:03 |
praefect | it is not easy to have this working on CentOS, griddynamics did a terrific job and their repo is now better than ever | 15:03 |
praefect | so it officialy works on centos with minimum hassle | 15:04 |
praefect | currently | 15:05 |
*** catintheroof has joined #openstack | 15:05 | |
Kiall | "<_rfz> -you need them on both" actually, you dont need both nova-network and nova-compute on every node.. it just makes sense to! | 15:10 |
_rfz | Kiall, what do you mean by making sense? :) | 15:12 |
*** andrewbogott has joined #openstack | 15:13 | |
_rfz | praefect, I added --auto_assign_floating_ip=True to nova.conf and it's not assigning floating point Ip's automatically | 15:13 |
Kiall | why have a single point of failure? combined with the overhead of "proxying" all traffic via a central nova-network | 15:13 |
*** supriya has joined #openstack | 15:13 | |
_rfz | Okay, so i did undertand it | 15:13 |
*** daMaestro has joined #openstack | 15:13 | |
Kiall | _rfz: if you're using my packages.. it should! I fixed that bug a while back | 15:13 |
_rfz | thanks | 15:13 |
*** hadrian has joined #openstack | 15:14 | |
daMaestro | What do I need to do to diagnose why I'm getting a 401 from object servers when doing a PUT but not a GET? | 15:14 |
daMaestro | I'm attempting to make a PUT be authorized by a signed URL via the swift3 interface. | 15:15 |
Kiall | sorry daMaestro, not a swift user :) | 15:15 |
*** cp16net has joined #openstack | 15:15 | |
_rfz | KIall, it's working now | 15:16 |
Kiall | _rfz: did you restart both network and compute after making the nova.conf change? | 15:16 |
*** cp16net has quit IRC | 15:16 | |
_rfz | I thought i did, but obviosly I didn't :) | 15:16 |
_rfz | it's working now | 15:16 |
*** cp16net has joined #openstack | 15:16 | |
_rfz | Kiall, BTW that bios error with libvirt went away on its own | 15:17 |
Kiall | on its own?? That doesnt make sense! | 15:17 |
Kiall | And is slightly worrying, since I cant remember how I fixed it either ;) | 15:17 |
_rfz | Hahaha yes | 15:18 |
*** PeteDaGuru has joined #openstack | 15:18 | |
*** daMaestro has quit IRC | 15:26 | |
andrewbogott | I switched my nova install from diablo to master, and now nova-api won't start up because of python dependency problems (at the moment, 'No module named metadatarequesthandler'). Can anyone suggest what packages I might need? | 15:27 |
andrewbogott | Or is there documentation someplace listing dependencies for master rather than just diablo? | 15:27 |
*** javiF has joined #openstack | 15:30 | |
*** rnirmal has joined #openstack | 15:34 | |
*** JStoker has quit IRC | 15:36 | |
*** JStoker has joined #openstack | 15:42 | |
soren | cloudgeek: It means we have the "voice" bit set. In an open channel, this doesn't have any technical meeting, but in this channel (and many others), it's used to denote people who are somehow core to the project. | 15:43 |
*** marrusl has quit IRC | 15:44 | |
Ramonster | :wq | 15:48 |
Ramonster | oops, wrong window sorry | 15:48 |
*** cmasseraf has joined #openstack | 15:51 | |
*** deshantm_laptop has quit IRC | 15:52 | |
*** dolphm has quit IRC | 15:54 | |
uvirtbot | New bug: #843069 in keystone "Tests failing on memcache backend (memcache testing disabled in trunk)" [High,Confirmed] https://launchpad.net/bugs/843069 | 15:56 |
uvirtbot | New bug: #894807 in nova "nova-manage vm list output is not consistent with nova cli - changes _ to - in name !" [Undecided,New] https://launchpad.net/bugs/894807 | 15:56 |
*** kaigan_ has quit IRC | 16:00 | |
*** mindpixel has quit IRC | 16:00 | |
uvirtbot | New bug: #894808 in nova "nova keypair-delete cannot cope with keypairs with complex names" [Undecided,New] https://launchpad.net/bugs/894808 | 16:01 |
*** dolphm has joined #openstack | 16:04 | |
*** reidrac has left #openstack | 16:04 | |
*** cereal_bars has joined #openstack | 16:04 | |
*** uksysadmin has quit IRC | 16:04 | |
*** troya has quit IRC | 16:06 | |
*** wi3dzma has quit IRC | 16:08 | |
*** freeflyi1g has joined #openstack | 16:11 | |
*** freeflying has quit IRC | 16:14 | |
*** guigui1 has quit IRC | 16:17 | |
*** katkee has quit IRC | 16:20 | |
*** dirkx has quit IRC | 16:23 | |
_rfz | Kiall, have you done run stress tests on your openstack environment? | 16:23 |
Kiall | Nothing major .. noticing any issues? | 16:25 |
Kiall | WormMan is probably the guy to talk to re stress testing ;) | 16:26 |
Kiall | I know he's done a fair bit of tuning... | 16:26 |
_rfz | I've started to run some tests now (creating 100vms) so far it's running okay | 16:27 |
_rfz | but the dashboard is real sloow | 16:27 |
Kiall | my default apache config is.. lacking.. | 16:27 |
Kiall | But, I believe thats an issue with the diablo dash anyway.. | 16:28 |
Kiall | the essex dashboard seems faster to kme | 16:28 |
Kiall | me* | 16:28 |
Kiall | (I've got the essex dash packaged up @ http://launchpad.net/~managedit/+archive/openstack-testing) | 16:28 |
Kiall | Its (mostly) compatible.. | 16:28 |
Kiall | and by mostly, I reckon they are just bugs in the essex dash | 16:28 |
andrewbogott | Hm... is anyone about who can advise me about using devstack with the nova master branch? Or is anyone about who knows about devstack, at all? | 16:32 |
*** yshh has quit IRC | 16:33 | |
*** rnorwood has joined #openstack | 16:33 | |
*** cloudgeek has quit IRC | 16:34 | |
*** nerens has joined #openstack | 16:37 | |
*** dobber has quit IRC | 16:40 | |
*** supriya has quit IRC | 16:41 | |
_rfz | Kiall, only 30 machines lauched of the 100 - and for the most part nova was unresponsive | 16:43 |
*** pradeep has quit IRC | 16:43 | |
*** cloudgeek has joined #openstack | 16:48 | |
dubenstein | «glance add» is taking too long, bursting machine load to maximum, have someone experienced an issue like that ? | 16:50 |
*** marcuz has quit IRC | 16:51 | |
*** zefry has joined #openstack | 16:52 | |
zefry | hello | 16:52 |
coli | _rfz: how did you lunch instances ? | 16:52 |
zefry | I have trouble on openstack | 16:53 |
zefry | can u help me?? | 16:53 |
zefry | anyone | 16:53 |
zefry | root@server1-cloud:/home/localadmin/creds# euca-describe-availability-zones verbose | 16:54 |
zefry | Warning: failed to parse error message from AWS: <unknown>:1:0: syntax error | 16:54 |
zefry | None: None | 16:54 |
zefry | hello anyone | 16:54 |
Kiall | _rfz: really? only 30? anything interesting in the logs? | 16:55 |
_rfz | coli, with a script that uses euca-run-instances | 16:55 |
_rfz | kiall, not much the compute service needed to be restarted a few times | 16:56 |
Kiall | no errors out of it? it just hung? | 16:56 |
Kiall | zefry: usually that means your EC2 credentials are incorrect | 16:57 |
Kiall | are you using keystone or not? | 16:57 |
coli | _rfz: do you run nova-api on seprate host ? did you do sleep X; between starting instances ? | 16:57 |
Kiall | zefry: and, how did you install nova? from ubuntu repos? devstack? something else? | 16:58 |
zefry | i'm install openstack on ubuntu server 11.10 | 16:58 |
_rfz | coli, I just run nova.api on the controller host, no sleep between creating the instances | 16:58 |
Kiall | zefry: and the other question re keystone? | 16:59 |
coli | _rfz: remaind me how many computing hosts ? | 16:59 |
coli | _rfz: remember that instances use nova-api as well when booting in order to get the metadata. | 17:00 |
zefry | Kiall: have a solutions?? | 17:00 |
coli | _rfz: did you try to spread the nova-appi onto compute hosts ? | 17:00 |
Kiall | zefry..maybe.. if you can answer the questions ;) | 17:00 |
_rfz | coli, I will try installing nova-api on the compute nodes - and let you know how it goes | 17:01 |
*** nerens has quit IRC | 17:02 | |
_rfz | Kiall, no errore just hangs | 17:03 |
Kiall | _rfz: yea.. the EC2 metadata part of nova-api is being split into a new service in essex i believe.. well, not a new service.. but something you can run a nova-api for just the MD | 17:03 |
*** dolphm has quit IRC | 17:05 | |
_rfz | kiall, when I try install nova-api on the compute node I get this error: ImportError: No module named keystone.middleware.ec2_token | 17:06 |
coli | kiall: it's garbing it from the central mysql so it's still a bottleneck however nova-api middleware is working on the compute hosts. I cannot see the reason for spliting it into a special part just for metadata, however I have seen some other blueprints for essex which don't make any sens from architectural and design point of view (just for interest of some comapnies) | 17:06 |
Kiall | _rfz: python-keystone is needed to | 17:06 |
Kiall | too* | 17:06 |
_rfz | cheers | 17:06 |
Kiall | if you use keystone, since the API paste references a keystone class for keystone auth | 17:07 |
cloudgeek | bye all | 17:07 |
coli | _rfz: you need python-keystone | 17:07 |
*** katkee has joined #openstack | 17:07 | |
*** vdo has quit IRC | 17:07 | |
*** chemikadze has quit IRC | 17:08 | |
*** chemikadze has joined #openstack | 17:09 | |
*** debo-os has quit IRC | 17:11 | |
*** dolphm has joined #openstack | 17:12 | |
*** fysa has joined #openstack | 17:14 | |
*** zefry- has joined #openstack | 17:17 | |
*** Ramonster has quit IRC | 17:21 | |
*** woleium has joined #openstack | 17:21 | |
*** _rfz has quit IRC | 17:21 | |
*** negronjl has quit IRC | 17:23 | |
*** omsynge has quit IRC | 17:23 | |
*** katkee has quit IRC | 17:26 | |
*** rnorwood has quit IRC | 17:26 | |
*** rnorwood has joined #openstack | 17:30 | |
*** hugokuo has joined #openstack | 17:34 | |
*** koolhead17 has joined #openstack | 17:38 | |
*** javiF has quit IRC | 17:38 | |
*** buzzy has joined #openstack | 17:38 | |
buzzy | I have problems using variables in Hive. I hope someone in here knows. I use something like hive -hiveconf test=hello -f /tmp/test.sql | 17:39 |
buzzy | but I cannot read the variable in the .sql file. I have tried with "set test; describe ${test}" | 17:39 |
*** shaon has joined #openstack | 17:41 | |
zykes- | hiveconf ? | 17:41 |
zykes- | what does that have to do with openstack buzzy ? | 17:41 |
koolhead17 | hola zykes- | 17:41 |
zykes- | ello | 17:41 |
koolhead17 | buzzy: yoor in wrong channel | 17:41 |
koolhead17 | winston-d: hells | 17:42 |
koolhead17 | hellos | 17:42 |
koolhead17 | zykes-: wassup | 17:42 |
zykes- | very little | 17:42 |
zykes- | we're concidering doing OS Swift @ Work | 17:42 |
zykes- | which is cool news | 17:42 |
buzzy | the #hadoop channel told me to ask in here :) | 17:43 |
zykes- | nothing here buzzy for that | 17:44 |
zykes- | ;) | 17:44 |
koolhead17 | ;-) | 17:44 |
Kiall | lol | 17:45 |
*** krow has joined #openstack | 17:46 | |
*** buzzy has left #openstack | 17:47 | |
*** oonersch has quit IRC | 17:50 | |
hugokuo | 12345... | 17:50 |
*** magg has joined #openstack | 17:51 | |
magg | hello | 17:51 |
* koolhead17 kicks hugokuo :P | 17:51 | |
* koolhead17 wonders where did razique disappeared :P | 17:51 | |
hugokuo | koolhead , did not get your mail …. | 17:51 |
*** darraghb has quit IRC | 17:52 | |
magg | im getting a Boot failed not a bootable disk, No bootable device | 17:53 |
magg | on the vnc console | 17:53 |
magg | im using an oneiric image from kiall scripts | 17:53 |
Kiall | magg: I had that once, seems qemu-nbd got itself it a twist. I never found the root cause, but a reboot fixed it | 17:54 |
magg | oh | 17:54 |
*** cereal_bars has quit IRC | 17:55 | |
*** GheRivero has joined #openstack | 17:56 | |
koolhead17 | hugokuo i have taken few days break from all openstack related brain storming :D | 17:56 |
shaon | koolhead17, :P | 17:56 |
*** catintheroof has quit IRC | 17:57 | |
*** rustam has quit IRC | 17:57 | |
*** andrewbogott has quit IRC | 17:58 | |
*** mikhail has joined #openstack | 18:00 | |
*** hallyn has joined #openstack | 18:03 | |
magg | kiall: no luck with the oneiric image i get the boot failed.... the lucid one hangs... the tty image works fine | 18:05 |
Kiall | dont suppose you rebooted? | 18:06 |
koolhead17 | hey mtaylor | 18:06 |
mtaylor | hi koolhead17 | 18:06 |
coli | magg: what do you mean oneiric failed ? what have failed ? | 18:06 |
magg | kiall: an instance reboot or machine reboot? | 18:07 |
Kiall | machine, its the only think I could find to get it out of that state. | 18:08 |
magg | coli: i get a Boot failed not a bootable disk, No bootable device when try to connect with vnc | 18:08 |
*** chomping has joined #openstack | 18:08 | |
magg | oh i did an instance reboot lol | 18:09 |
*** foexle has quit IRC | 18:09 | |
Kiall | From memory, qemu-nbd --attach (or was it kpartx?) was failing, Nothing I did would let me succeed | 18:09 |
*** cereal_bars has joined #openstack | 18:09 | |
*** rnorwood has quit IRC | 18:09 | |
Kiall | (I wasn't even able to mount the image manually without nova...) | 18:09 |
*** catintheroof has joined #openstack | 18:10 | |
*** chomping has left #openstack | 18:10 | |
*** jedi4ever has quit IRC | 18:17 | |
*** adjohn has joined #openstack | 18:17 | |
*** egant has joined #openstack | 18:18 | |
magg | nope that didn't do it | 18:22 |
Kiall | magg: humm those images do work, maybe its a a bad download? | 18:23 |
Kiall | or a bad upload to glance, or a bad download from glance to nova | 18:23 |
magg | ...damn | 18:23 |
Kiall | (There is basically no error checking in all that I believe) | 18:23 |
Kiall | anyway, I'm off! cyas | 18:28 |
*** koolhead17 has quit IRC | 18:30 | |
*** CaptTofu has quit IRC | 18:31 | |
*** TheOsprey has joined #openstack | 18:33 | |
*** Ryan_Lane has joined #openstack | 18:33 | |
*** map_nw_ has quit IRC | 18:35 | |
*** andreacloud has joined #openstack | 18:35 | |
*** obino has quit IRC | 18:36 | |
*** map_nw has joined #openstack | 18:36 | |
magg | to bale to ping the instance or connect with ssh do i need to use euca tools | 18:37 |
magg | to be able** | 18:37 |
*** pasm has joined #openstack | 18:41 | |
pasm | andreacloud: ciao andrea | 18:41 |
magg | cuz on the compute node i can't ping or connect to the instance | 18:41 |
*** andreacloud has left #openstack | 18:41 | |
*** pasm has left #openstack | 18:42 | |
*** koolhead17 has joined #openstack | 18:42 | |
*** Ryan_Lane has quit IRC | 19:00 | |
*** llang629 has joined #openstack | 19:01 | |
*** cereal_bars has quit IRC | 19:02 | |
*** redconnection has quit IRC | 19:02 | |
*** nerens has joined #openstack | 19:03 | |
*** fysa has quit IRC | 19:03 | |
magg | ? help | 19:04 |
*** llang629 has left #openstack | 19:04 | |
*** krow has quit IRC | 19:06 | |
magg | i have this rule on the dashboard | 19:11 |
magg | tcp 22 22 0.0.0.0/0 | 19:11 |
magg | but i cant connect with ssh | 19:11 |
magg | anybody? | 19:11 |
*** arBmind has joined #openstack | 19:12 | |
*** _rfz has joined #openstack | 19:12 | |
*** cereal_bars has joined #openstack | 19:15 | |
*** nati2_ has quit IRC | 19:15 | |
magg | or how do i use euca-tools with keystone | 19:15 |
*** nati2 has joined #openstack | 19:16 | |
*** andrewbogott has joined #openstack | 19:18 | |
*** dubenstein has quit IRC | 19:21 | |
*** obino has joined #openstack | 19:23 | |
*** bhall has quit IRC | 19:26 | |
coli | magg: check iptables on the host with that VM and see if it's got right rules, nova-network during creation of an instance is creating appropirate rules in iptables based on security groups | 19:28 |
magg | kk | 19:29 |
magg | coli: http://pastebin.com/jNetJipU | 19:38 |
*** cmasseraf has quit IRC | 19:40 | |
*** hallyn has quit IRC | 19:43 | |
*** bengrue has joined #openstack | 19:46 | |
*** shaon has quit IRC | 19:46 | |
*** tjikkun has quit IRC | 19:56 | |
*** bhall has joined #openstack | 19:59 | |
coli | magg: did you analyse these iptables rules ? | 19:59 |
magg | coli: yeah | 20:02 |
*** hugokuo has quit IRC | 20:02 | |
magg | on the filter table it adds a rule for tcp and udp on port 22 | 20:02 |
*** pixelbeat has quit IRC | 20:04 | |
*** PotHix has quit IRC | 20:19 | |
magg | ok i got it to work | 20:28 |
*** arBmind has quit IRC | 20:32 | |
*** magg has quit IRC | 20:41 | |
*** mnabil has joined #openstack | 20:43 | |
*** TheOsprey has quit IRC | 20:43 | |
*** adjohn has quit IRC | 20:50 | |
*** adjohn has joined #openstack | 20:50 | |
*** adjohn has quit IRC | 20:50 | |
*** Ryan_Lane has joined #openstack | 20:54 | |
*** catintheroof has quit IRC | 20:57 | |
Ryan_Lane | ugh | 21:02 |
Ryan_Lane | in diablo users still need to be in both a global and a project group to be considered to be in a group? | 21:02 |
*** praefect has quit IRC | 21:02 | |
*** n8 has joined #openstack | 21:04 | |
*** n8 is now known as Guest12154 | 21:04 | |
*** chemikadze has quit IRC | 21:05 | |
*** chemikadze has joined #openstack | 21:06 | |
*** cereal_bars has quit IRC | 21:11 | |
*** adjohn has joined #openstack | 21:14 | |
*** bhall has quit IRC | 21:22 | |
*** Guest12154 has quit IRC | 21:25 | |
*** jaypipes has joined #openstack | 21:29 | |
*** mikhail has quit IRC | 21:30 | |
*** negronjl has joined #openstack | 21:32 | |
*** troya has joined #openstack | 21:33 | |
*** bhall has joined #openstack | 21:38 | |
*** Ryan_Lane has quit IRC | 21:39 | |
*** magg has joined #openstack | 21:41 | |
*** mikhail has joined #openstack | 21:43 | |
*** n8 has joined #openstack | 21:43 | |
*** n8 is now known as Guest78795 | 21:44 | |
*** tjikkun has joined #openstack | 21:45 | |
*** tjikkun has joined #openstack | 21:45 | |
*** Ryan_Lane has joined #openstack | 21:45 | |
*** dolphm has quit IRC | 21:47 | |
*** magg has quit IRC | 21:49 | |
*** magg has joined #openstack | 21:50 | |
magg | hey kiall: are u sure the oneiric image is working | 21:50 |
magg | the lucid one and the tty one are though | 21:50 |
magg | i mean i got it to boot | 21:51 |
magg | i get a grub | 21:51 |
magg | but doesnt start | 21:51 |
magg | something is fishy about that image | 21:51 |
magg | i can smell it | 21:51 |
*** gohko_nao has quit IRC | 21:52 | |
Kiall | magg, so lucid and tty boot? | 21:54 |
Kiall | and oneiric doesnt? or? | 21:54 |
*** gohko_nao has joined #openstack | 21:55 | |
*** GheRivero has quit IRC | 21:55 | |
*** hingo has quit IRC | 22:03 | |
*** Ryan_Lane has quit IRC | 22:05 | |
magg | yeah | 22:07 |
magg | oneiric does boot | 22:07 |
magg | i the grub screen | 22:07 |
magg | i get** | 22:07 |
magg | but doesnt start | 22:07 |
magg | brb | 22:07 |
*** dolphm has joined #openstack | 22:11 | |
*** magg has quit IRC | 22:11 | |
*** magg has joined #openstack | 22:12 | |
magg | back | 22:13 |
*** hky has joined #openstack | 22:13 | |
*** Guest70665 has joined #openstack | 22:14 | |
magg | so kiall? | 22:20 |
*** debo-os has joined #openstack | 22:20 | |
*** Ryan_Lane has joined #openstack | 22:22 | |
*** rnirmal has quit IRC | 22:23 | |
*** andrewbogott has quit IRC | 22:24 | |
*** mattstep has joined #openstack | 22:24 | |
*** bengrue has quit IRC | 22:25 | |
*** Guest70665 has quit IRC | 22:26 | |
*** dolphm has quit IRC | 22:27 | |
*** afm has joined #openstack | 22:32 | |
afm | evening all. following the openstack docs for the first time.. running 'euca-register mybucket/server.img.manifest.xml' returns UnknownError: An unknown error has occurred. Please try your request again. | 22:33 |
afm | tried from scratch twice now… can't seem to get the image and or bundle registered | 22:33 |
*** cloudgeek has quit IRC | 22:34 | |
*** admgre has joined #openstack | 22:39 | |
afm | Uploaded image as images/ubuntu-11_x64.img.manifest.xml | 22:39 |
afm | root@ubuntu:~# euca-register images/ubuntu-11_x64.img.manifest.xml | 22:39 |
afm | UnknownError: An unknown error has occurred. Please try your request again. | 22:39 |
*** dolphm has joined #openstack | 22:39 | |
*** magg has quit IRC | 22:44 | |
*** bengrue has joined #openstack | 22:45 | |
*** cloudgeek has joined #openstack | 22:47 | |
*** bengrue has quit IRC | 22:54 | |
*** mnabil has quit IRC | 23:01 | |
*** PeteDaGuru has quit IRC | 23:08 | |
*** krow has joined #openstack | 23:09 | |
*** matiu_ has joined #openstack | 23:13 | |
*** matiu_ has joined #openstack | 23:13 | |
*** mnabil has joined #openstack | 23:13 | |
*** matiu_ has joined #openstack | 23:14 | |
*** matiu_ has joined #openstack | 23:14 | |
*** rsampaio has quit IRC | 23:15 | |
*** matiu_ has joined #openstack | 23:15 | |
*** matiu_ has joined #openstack | 23:15 | |
*** adjohn has quit IRC | 23:17 | |
*** matiu_ has quit IRC | 23:19 | |
*** debo-os has quit IRC | 23:20 | |
*** adjohn has joined #openstack | 23:21 | |
*** adjohn has quit IRC | 23:21 | |
*** mikhail has quit IRC | 23:22 | |
*** arBmind has joined #openstack | 23:23 | |
*** rustam has joined #openstack | 23:26 | |
*** nerens has quit IRC | 23:31 | |
*** nati2 has quit IRC | 23:31 | |
*** nati2 has joined #openstack | 23:33 | |
*** troya has joined #openstack | 23:34 | |
duffman | Is it difficult to migrate from proxmox to openstack? | 23:46 |
*** Ryan_Lane has quit IRC | 23:49 | |
*** ryan_fox1985 has quit IRC | 23:54 | |
*** Guest78795 has quit IRC | 23:56 | |
*** HugoKuo__ has joined #openstack | 23:57 | |
*** magg has joined #openstack | 23:59 | |
magg | yo! | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!