The no-name cloud at DOC has the following catchy address: 146.169.37.1
(1) Get an account
https://146.169.37.1:8443
(2) euca2ools software
Still installed from accessing our local cloud.
(3) Credentials
This follows the instructions here
As me on deathstar:
https://cloud00.grid.hep.ph.ic.ac.uk:8443/#credentials
"Download Credentials"
This makes euca2-dbauer-x509.zip
It's not immediately obvious to me how I am going to manage the credentials
for two different clouds, so for the time being, I move
.euca/euca2-dbauer-x509 to .euca/euca2-dbauer-x509-cloud00 and put the new
euca2-dbauer-x509 from DOC in its place.
[As it turns out, I only need to source
. ~/.euca/euca2-dbauer-x509[-cloud00]/eucarc to get the setting for the
appropriate cloud.]
unzip the file:
ark euca2-dbauer-x509.zip
For now I am just going to stick with Te permissions that worked last
time:
chmod 0700 .euca/euca2-dbauer-x509
chmod 0600 ~/.euca/euca2-dbauer-x509/*
(4) "Hello World", using a preloaded image
(4a) Setup the environment and check what's there:
(At this point, I really really need to use bash ;-)
. ~/.euca/euca2-dbauer-x509/eucarc
[dbauer@deathstar ~]$ euca-describe-images
IMAGE eki-89ED1387 lucid-kvm/vmlinuz-2.6.35-25-virtual.manifest.xml
admin available public x86_64 kernel
instance-store
IMAGE eri-25A61539
cernvm-images/initrd-2.6.18-194.11.4.el5.img.manifest.xml admin
available public x86_64 ramdisk instance-store
IMAGE eki-EE69145D cernvm-images/vmlinuz-2.6.18-194.11.4.el5.manifest.xml
admin available public x86_64 kernel
instance-store
IMAGE eri-C5BC1468 lucid-kvm/initrd.img-2.6.35-25-virtual.manifest.xml
admin available public x86_64 ramdisk
instance-store
IMAGE emi-F6A8157B lucid-kvm/ubuntu-lucid-kvm-doc_1-4.img.manifest.xml
admin available public x86_64 machine eki-89ED1387
eri-C5BC1468 instance-store
IMAGE emi-5E041712 cernvm-images/cernvm-2.2.0-kvm.img-fulldisk.manifest.xml
admin available public x86_64 machine eki-EE69145D eri-25A61539
instance-store
[dbauer@deathstar ~]$ euca-describe-instances
[nothing]
[dbauer@deathstar ~]$ euca-describe-availability-zones
AVAILABILITYZONE lesc-zone-2 10.53.0.1
(4b) Get a keypair:
[dbauer@deathstar ~]$ euca-add-keypair lesc00 | tee lesc00.private
chmod 0600 lesc00.private
[dbauer@deathstar ~]$ euca-describe-keypairs
(should this be public ? Maybe I'll slightly modify it ....)
KEYPAIR lesc00 99:65:51:72:34:66:0a:04:3f:35:53:e5:6f:73:fd:d1:df:c1:4e:de
(4c) Go for it:
[dbauer@deathstar ~]$ euca-run-instances -k lesc00 -n 1 emi-5E041712
VmTypeVerify: image too large [size=6267MB] for instance type m1.small
[disk=5120MB]
or not as it happens ...
[dbauer@deathstar ~]$ euca-run-instances -k lesc00 -n 1 -t m1.large emi-5E041712
RESERVATION r-4D28088B dbauer dbauer-default
INSTANCE i-3BE80724 emi-5E041712 0.0.0.0 0.0.0.0 pending lesc00 0 m1.large 2011-05-05T12:00:59.711Z lesc-zone-2 eki-EE69145D eri-25A61539
[dbauer@deathstar ~]$ euca-describe-instances
RESERVATION r-4D28088B dbauer default
INSTANCE i-3BE80724 emi-5E041712 146.169.37.105 10.40.1.66
pending lesc00 0 m1.large 2011-05-05T12:00:59.711Z
lesc-zone-2 eki-EE69145D eri-25A61539
Wait for it to change to running .... and wait ... and wait ....
Once it's up and running:
[dbauer@deathstar ~]$ ssh -i lesc00.private root@146.169.37.105
ssh: connect to host 146.169.37.105 port 22: Connection timed out
Go and open a port:
[dbauer@deathstar ~]$ euca-authorize -P tcp -p 22 -s 155.198.211.22 default
default None None tcp 22 22 155.198.211.22
Groups: 1
Nope.
Now try the subnet:
[dbauer@deathstar ~]$ euca-authorize -P tcp -p 22 -s 155.198.211.0/24 default
default None None tcp 22 22 155.198.211.0/24
GROUP default
PERMISSION default ALLOWS tcp 22 22 FROM CIDR 155.198.211.0/24
[dbauer@deathstar ~]$ ssh -i lesc00.private root@146.169.37.105
root@146.169.37.105's password:
After some further clarification it turns out the CERNVM is not configured to
obtain the user key from the Eucalyptus meta-data service on boot.
(4 d) Detour: Try the stock ubuntu image at DOC
[dbauer@deathstar ~]$ euca-run-instances -k lesc00 -n 1 emi-F6A8157B
[dbauer@deathstar ~]$ euca-describe-instances
[snip]
INSTANCE i-47E508B7 emi-F6A8157B 146.169.37.106 10.40.1.67
running lesc00 0 m1.small 2011-05-05T14:21:08.652Z
lesc-zone-2 eki-89ED1387 eri-C5BC1468
ssh -v -i lesc00.private root@146.169.37.106
Welcome to Ubuntu!
* Documentation: https://help.ubuntu.com/
root@ip-10-40-1-67:~# echo "hello world"
hello world
root@ip-10-40-1-67:~# exit
[dbauer@deathstar ~]$ euca-terminate-instances i-47E508B7
(4 e)Now back to that CERNVM
I need to use the web interface,
here are the
instructions.
This is relatively straight forward - once I've run euca-authorize -P tcp -p 8003 -s 155.198.211.0/24 default.
I set the admin password and then
create a user 'dbauer', selecting 'cms' and /bin/bash.
I change the (automatically seletced) site name from T3_UK_London_RHUL to
T2_UK_London_IC, which might or might not have been a good idea: See here.
I can now ssh into the machine:
[dbauer@deathstar ~]$ ssh dbauer@146.169.37.105
dbauer@146.169.37.105's password:
Last login: Mon May 9 14:40:14 2011 from deathstar.hep.ph.ic.ac.uk
[dbauer@host-10-40-1-66 ~]$
and as the 'first' user I am allowed to run sudo:
[dbauer@host-10-40-1-66 ~]$ sudo updatedb
[sudo] password for dbauer:
Success
Next step: Running CMS MC on the cloud.