[01:09] <smoser> kiall, there should not be, but giving console output (euca-get-console-output | pastebin ) would be helpful.  also, what version of "ubuntu instance"
[01:11] <kiall> smoser, 10.10 UEC and 10.10 instance, will grab the console output now ..
[01:11] <liam> i need to connect to my ubuntu ec2 server from my ubuntu machine using filezilla with sftp but it won't connect. How can I specify the key to use in filezilla?
[01:11] <kiall> console output: http://pastebin.com/RHkixNnS
[01:12] <kiall> liam, they have a howto for that :) http://wiki.filezilla-project.org/Howto
[01:14] <liam> kiall: it wants a .ppk..... do you know how I can convert my ec2.pem to a .ppk?
[01:15] <kiall> can puttygen (PPK is Putty's key format) not do that?
[01:15] <kiall> (Its been a while since I've used windows...)
[01:16] <kiall> yea it can .. http://invoice.geeksontheway.com/wiki/index.php/Working_with_EC2#Converting_pem_files_to_ppk_files_for_use_with_putty
[01:37] <kiall> smoser, just noticed the " | pastebin" ... how didnt I know about the pastebinit package before ..  :)
[09:51] <kevinw> can anyone shed light on the error: EXPIRED: Instance expired after not being reported for 600000 ms
[11:19] <kevinw> also secondly, if both the node controller and the cloud controller are both on public routable IP's, which interface do i bridge to make the VM accessible with a public routable IP?
[12:18] <TeTeT> kevinw: the cloud controller/front end interface. If you check the iptables rules there, you see a NAT for the public IP for any instance you've setup
[12:19] <kevinw> TeTet: im not doing any nat
[12:20] <TeTeT> kevinw: it's automatically done on the front end if your instances is using a public / private network address
[12:27] <kevinw> ok, ill look into that, my main problem is that the machines wont start, they stay as pending then come up with an error about not being reported in 600000ms
[12:32] <TeTeT> kevinw: you want to check the node controllers log files, e.g. /var/log/eucalyptus/nc.log
[12:32] <TeTeT> do a ''tail -f /var/log/eucalyptus/nc.log | grep -v doDescribe'' when starting an instance and watch for any errors
[12:34] <kevinw> root@node1:/var/log/eucalyptus# tail -f /var/log/eucalyptus/nc.log | grep -v doDescribe
[12:34] <kevinw> [Mon Dec  6 12:33:10 2010][001197][EUCADEBUG ] doStartNetwork() invoked
[12:34] <kevinw> [Mon Dec  6 12:33:10 2010][001197][EUCAINFO  ] StartNetwork(): SUCCESS return from vnetStartNetwork 0
[12:34] <kevinw> [Mon Dec  6 12:33:10 2010][001197][EUCAINFO  ] StartNetwork(): done
[12:34] <kevinw> it gets an IP ok but then just sits pending
[12:36] <TeTeT> kevinw: hmm, is the front end's walrus service reachable from the node? What's the output of sudo euca_conf --list-walruses
[12:37] <kevinw> root@node1:/var/log/eucalyptus# sudo euca_conf --list-walruses
[12:37] <kevinw> ERROR: cannot locate eucalyptus database, try logging in through the admin web interface.
[12:37] <kevinw> root@node1:/var/log/eucalyptus#
[12:37] <kevinw> root@cloudcc1:/var/log/eucalyptus# sudo euca_conf --list-walruses
[12:37] <kevinw> registered walruses:
[12:37] <kevinw>    walrus  213.48.214.61
[12:38] <kevinw> how do i go about registering walruses on the node controller?
[12:39] <kevinw> scrap that, got that working
[12:39] <kevinw> root@cloudcc1:/var/log/eucalyptus# sudo euca_conf --list-walruses
[12:39] <kevinw> registered walruses:
[12:39] <kevinw>    walrus  213.48.214.61
[12:41] <TeTeT> kevinw: can you ping 213.48.214.61 from the node controller, is it reachable?
[13:08] <kevinw> yep,
[13:08] <kevinw> pings between the two physical servers are OK
[14:17] <kevinw> the instances stay as pending no matter what I do
[14:18] <TeTeT> kevinw: is there ever a kvm started on your node controller? Is the image copied from walrus to the local storage?
[14:18] <kevinw> how do i check that?
[14:19] <TeTeT> kevinw: check /var/lib/eucalyptus/instances/admin on the node controller
[14:20] <TeTeT> kevinw: are there any directories named 'i-<id>'?
[14:20] <TeTeT> kevinw: and inside there should be a name called 'disk'
[14:20] <kevinw> nope there isnt
[14:21] <TeTeT> hmm
[14:21] <TeTeT> kevinw: is there a cache directory, /var/lib/eucalyptus/instances/eucalyptus/cache with emi- and eki- inside?
[14:21] <kevinw> theres a directory but its empty
[14:23] <kevinw> if it helps im not using the admin account,
[14:25] <TeTeT> kevinw: somethings really strange, can you try as admin to rule out any problem with the user? Not that there should be one
[14:28] <TritoLux> With UEC 10.10, Is there any reason why iscsi volumes get unaccessible after a SC reboot? Everything is working fine but if the server gets restarted then all existing volumes are not accessible anymore. Did anybody managed to get EBS to work fine on 10.10?
[14:50] <kevinw> TeTeT - same nothing in either directory
[14:51] <kevinw> should the walrus be registered on both the cloud controller and the node controller?
[14:51] <TeTeT> kevinw: nope, the front end is enough.
[14:51] <kevinw> the cc?
[14:52] <TeTeT> yes
[14:52] <kevinw> hmm
[14:52] <kevinw> the instances have IP's from the allocated range OK so that shows comms between the servers is OK,
[14:52] <kevinw> i wonder why its not registering
[14:52] <TeTeT> hmm, when you do a euca-describe-images, do you see the uploaded images? Is there something in /var/lib/eucalyptus/bukkits?
[14:53] <TeTeT> somehow the image from walrus fails to be transfered to your node controller
[14:53] <TeTeT> something in, meant something in on the front end
[14:54] <kevinw> root@cloudcc1:/var/lib/eucalyptus/bukkits# ls
[14:54] <kevinw> image-store-1291376044  image-store-1291393809  image-store-1291393871
[14:54] <kevinw> im using the images from the store
[14:57] <TeTeT> ok, when you copy your eucarc to the node controller, source it there, can you do an euca-bundle-image /etc/hosts; euca-upload-bundle -m /tmp/<manifest> -b etc-hosts?
[14:57] <TeTeT> wonder if the upload works
[15:09] <kevinw> ill give it a try, i notice in the logs that its trying to get to http://213.48.214.62:8775/axis2/services/EucalyptusNC but I cant get to that in a browser..
[15:10] <TeTeT> kevinw: hmm, seems your network setup between front-end and node controller is no good then
[15:12] <kevinw> i agree, but i dont see where, its two servers, with 1 NIC each, plugged into a switch..
[15:12] <TeTeT> is there a firewall running on either NC or front-end that might block the traffic?
[15:14] <kevinw> only whatever was installed by the setup
[15:14] <smoser> erichammond, awake ?
[15:15] <TeTeT> kevinw: hmm, then there's nothing there. did the upload to walrus eventually work?
[15:16] <smoser> kiall, i would guess that you have something in /etc/fstab that mountall is waiting on
[15:17] <kevinw> im going to scrap and start again, for clarification theres no reason why have the CC and the NC on public IP's with the VM on the same range wont work
[15:17] <smoser> well, erichammond if you do see this, alestic.com gives me: "Gateway Time-out
[15:17] <smoser> The gateway did not receive a timely response from the upstream server or application." when i try to login with my launchpad id (https://launchpad.net/~smoser)
[15:25] <CharlieSu> How do you change limits like this on an Ec2 instance? (*               soft    nofile           65536)  do you have to edit limits.conf and restart the server?
[15:49] <jmgalloway> is it possible to create a vm without exchanging the key?
[17:34] <CharlieSu> How do you change limits like this on an Ec2 instance? (*               soft    nofile           65536)  do you have to edit limits.conf and restart the server?
[19:03] <erichammond> smoser: Just got up, but need to commute for hour, then scheduled lunch.
[19:03] <smoser> fair. i had this bug once before when trying launchpad openid on alestic.com also
[19:03] <smoser> ie, its not new, but i'd like to use my  launchpad.net id
[19:03] <erichammond> smoser: I saw a gateway timeout on http://alestic.com last night, but refresh worked.
[19:04] <smoser> tried it twice... no dice, but i can try again
[19:05] <erichammond> smoser: I get it regularly when using dynamic MT (Movable Type) pages.
[19:05] <erichammond> Thought it was a bug in MT, but it might instead be related to a DDOS defense tweak I made a while back for another web site on that server.
[19:05] <erichammond> I'll look into it later today.
[19:05] <smoser> well i consistently get it.
[19:05] <erichammond> hm
[19:05] <smoser> you can try too, just try your launchpad account
[19:06] <erichammond> ah, yes, consistent with MT dynamic stuff.
[19:06] <erichammond> especially if it is requiring a wait on an external resource like launchpad.
[21:33] <erichammond> smoser: I tweaked timeouts on http://alestic.com
[21:34] <erichammond> smoser: Let me know if you still have problems logging in.
[22:28] <asac> odd ... a few hours ago i was able to ping something like www.yahoo.com from my ec2 instance ... doesnt work anymore ... i am trying to stop the instance in web interface ... and that doesnt work either it seems :/ ... it stays in stopping state and i can still ssh into it
[22:28] <asac> should i just run sudo halt ;)?
[22:33] <asac> crazy
[22:33]  * asac feels lost... and scared ... not even http://us-east-1.ec2.archive.ubuntu.com is reachable anymore
[22:33] <asac> but i can ssh in !!!