=== da is now known as davinci11 [06:28] can't start instance and get this output [ FinishedVerify: Not enough resources (0 < 1: vm instances.Not enough resources (0 < 1: vm instances.] [06:49] davinci11: if you are using euca2ools, what is the output of euca-describe-availability-zones verbose? [06:50] AVAILABILITYZONE |- vm types free / max cpu ram disk [06:50] AVAILABILITYZONE |- m1.small 0000 / 0000 1 192 2 [06:50] AVAILABILITYZONE |- c1.medium 0000 / 0000 1 256 5 [06:50] AVAILABILITYZONE |- m1.large 0000 / 0000 2 512 10 [06:50] AVAILABILITYZONE |- m1.xlarge 0000 / 0000 2 1024 20 [06:50] AVAILABILITYZONE |- c1.xlarge 0000 / 0000 4 2048 20 [06:50] so that means your compute nodes aren't registered within the cluster [06:51] i tried to deregister and register nodes again [06:51] but still same [06:51] look at euca_conf --list-nodes [06:52] no output [06:52] :( [06:52] maybe try removing the cluster and re-adding it, then adding the nodes? [06:53] also check the keys [06:53] tried as well too [06:54] even remove the key and created new one [06:55] sorry, that's about as much as I know. I got Eucalyptus working after reinstalling a few times. I remember having issues if the auto-discover didn't work during the uec install. [06:57] no other way without reinstall ?? :('' [06:59] Maybe someone with more eucalyptus experience can help you, I'm not familiar enough with it. I assume you've checked the config files to make sure the nodes are looking at the right address, and that the correct ports are open? [07:00] look at nodes machine ? or cc [07:00] maybe tits literal [07:00] not enough resources [07:00] try googling the error [07:01] already tried what i found [07:01] go to the compute nodes and check you can connect to the right port on the CC using telnet [07:03] i'm using ssh to CC is that means ok? [07:03] ssh goes over port 22, you want to check the node can connect to the CC over a different port that eucalyptus is using [07:04] how to check ? [07:05] On the front-end, ports 8443, 8773, 8774 and 9001 must be open; on the node, port 8775 must be open. [07:06] so on the CC, try telnet [IP-address-of-node] 8775 [07:06] if the connection is accepted then the port is open and the eucalyptus compute node is ok === kooolhead11|afk is now known as kooolhead11 [07:07] then SSH to the compute node, try to telnet [IP-address-of-CC] 8773 [07:07] and check the other numbers as well. [07:07] cant telnet from CC to nodes [07:07] port 8775 [07:09] check if the eucalyptus process is running on the nodes - SSH to the node and run ps aux | grep euca === ubc1_ is now known as davin11 === davin11 is now known as davinc1 [07:13] root 1824 0.0 0.1 5108 2148 ? Ss 12:22 0:00 apache2 -f /var/run/eucalyptus/httpd-nc.conf -D FOREGROUND [07:13] 107 1913 0.0 0.2 40048 4748 ? Sl 12:22 0:01 apache2 -f /var/run/eucalyptus/httpd-nc.conf -D FOREGROUND [07:13] root 2071 0.0 0.0 2352 912 ? Ss 12:22 0:00 avahi-publish -s cluster1 node _eucalyptus._tcp 8775 txtvers=1 protovers=1.5.0 type=node [07:13] ubc1 23024 0.0 0.0 3608 868 pts/1 S+ 15:09 0:00 grep --color=auto euca [07:13] ok [07:13] so the process is running [07:14] try telnet localhost 8775 [07:14] on the compute node [07:14] Trying 127.0.0.1... [07:14] Connected to localhost. [07:14] Escape character is '^]'. [07:14] ok [07:15] escape from that [07:15] so I'm not 100% sure, but it seems like your CC can't connect to your NC [07:15] which is usually a firewall issue [07:16] portmap will check the port satus [07:16] *status sorry [07:18] Chain INPUT (policy ACCEPT) [07:18] target prot opt source destination [07:18] ACCEPT udp -- anywhere anywhere udp dpt:domain [07:18] ACCEPT tcp -- anywhere anywhere tcp dpt:domain [07:18] ACCEPT udp -- anywhere anywhere udp dpt:bootps [07:18] ACCEPT all -- 192.168.122.0/24 anywhere [07:18] ACCEPT all -- anywhere anywhere [07:18] REJECT all -- anywhere anywhere reject-with icmp-port-unreachable [07:18] REJECT all -- anywhere anywhere reject-with icmp-port-unreachable [07:18] iptables from CC [07:21] I'm not sure...sorry [08:12] davinc1: maybe your cloud is not setup correctly? what's the output of euca-describe-availability-zones verbose? [08:13] AVAILABILITYZONE |- vm types free / max cpu ram disk [08:13] AVAILABILITYZONE |- m1.small 0000 / 0000 1 192 2 [08:13] AVAILABILITYZONE |- c1.medium 0000 / 0000 1 256 5 [08:13] AVAILABILITYZONE |- m1.large 0000 / 0000 2 512 10 [08:13] AVAILABILITYZONE |- m1.xlarge 0000 / 0000 2 1024 20 [08:13] AVAILABILITYZONE |- c1.xlarge 0000 / 0000 4 2048 20 [08:14] it work before, but now can start anymore [08:15] davinc1: you seee the 0000, that means your node controllers are not in the cloud anymore [08:15] davinc1: did you reboot the systems or something like that? [08:15] yeah. both CC and NC rebooted [08:15] still same [08:18] davinc1: from the front-end, do a $ sudo euca_conf --list-nodes [08:19] no output [08:19] davinc1: so for whatever reasons your node controller was dropped from eucas config [08:20] davinc1: try a $ sudo euca_conf --discover-nodes [08:20] and see if the node controller is picked up again [08:21] davinc1: what Ubuntu version do you use btw? [08:21] INFO: We expect all nodes to have eucalyptus installed in //var/lib/eucalyptus/keys for key synchronization. [08:21] output of $ sudo euca_conf --discover-nodes [08:22] davinc1: weird, what would remove the keys from there on the front-end? I'm a bit out of ideas right now [08:22] Ubuntu Enterprise Cloud (UEC) 11.04 [08:23] davinci11: can you ping the node controller from the front-end? maybe $ sudo euca_conf --register-nodes will work? [08:24] yep both work [08:24] davinci11: try the euca-describe-availability-zones verbose again [08:25] same [08:25] and euca_conf --list-nodes? [08:26] still no output [08:27] davinci11: what's the oputput of $ sudo euca_conf --register-nodes ? [08:28] INFO: We expect all nodes to have eucalyptus installed in //var/lib/eucalyptus/keys for key synchronization. [08:28] Trying rsync to sync keys with "192.168.168.20"...done. [08:30] davinci11: that looks ok, when you do a '$ sudo su - eucalyptus' on the front-end [08:30] davinc11: and then ssh to the node controller, does that work? [08:32] yes is working [08:33] davinci11: looks like a bug to me then. Daviey, do you have an opinion on the above node registration problem on 11.04? [09:04] davinci11: maybe this bug is hitting you: https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/585108 [09:04] Launchpad bug 585108 in eucalyptus "euca_conf --register-nodes returns 0 but the node is not registered" [Medium,Confirmed] [09:04] davinci11: probably dereigstering the node and then trying to add it again will help, but I'm not sure [09:07] TeTeT: I tried that way before, but still not work [09:09] davinci11: hmm, on the front-end, try a $ sudo restart eucalyptus CLEAN=1 [09:09] davinci11: and then try the registering dance again [09:13] still same === daker_ is now known as daker === F000 is now known as F00D === F00D is now known as F000 === dendro-afk is now known as dendrobates === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates === dendrobates is now known as dendro-afk [17:41] kim0, great blog post today. i'm sitting in a hadoop class right now. nice timing. [17:41] marrusl: hehe nice [17:41] marrusl: is canonical supporting hadoop professionally now [17:43] kim0, for premium customers I would think so yes. but probably not in general unless it went into main. [17:43] got it [17:54] hi all === daker is now known as daker_ === dendro-afk is now known as dendrobates === skrewler_ is now known as skrewler === dendrobates is now known as dendro-afk === niemeyer_ is now known as niemeyer === dendro-afk is now known as dendrobates === erichammond1 is now known as erichammond === erichammond1 is now known as erichammond === dendrobates is now known as dendro-afk