setting Apollo state to FinalProvisioning...
Goto your apollo dev envirenment.
https://apollo.amazon.com/environments/DevDesktop/ripwang/-
add your host. eg. yourname.desktop.amazon.com to the hosts and hostclasses, it will show "provisioning"...
wait for a while when it goes further.
Still failed at "Ack!infra-svcs did not return any hosts records! ..... sleep for *h**m**s - - ENTER (or SIGHUP) to retry",
Trying the way provided by IND
nce entering into the single user mode, do the following.
1. mount -o remount,rw /
2. sudo /usr/bin/chattr �Ci /etc/resolv.conf
3. sudo /bin/vi /etc/resolv.conf
Add following DNS names
nameserver 10.45.65.42
nameserver 10.43.120.49
nameserver 10.48.65.33
nameserver 10.4.4.10
4. sudo /usr/bin/chattr +i /etc/resolv.conf
5. make a backup file: cp /etc/lwresd.conf /etc/lwresd.conf1234
edit following lines:
forwarders { 10.45.65.42; 10.43.120.49; 10.48.65.33; 10.4.4.10; }
6. Restart the system and login into default mode
Please dont forget to delete the old DNS servers starting with 172.25.XXXXX. when updating in both resolve.conf and lwresd.conf
It prompts to enter web-indentity again, and it seems positive....
now the build had been preliminary build, but failed to log in......
what I tried:
1, https://security-bastions.amazon.com/ProdBastions/RHEL5/
Tools
Run this command:
sudo yum install amazon-security-bastion-prod
Once the RPM has finished installing, run this command (as yourself, no need for SUDO):
/usr/local/bin/security-bastion-setup.sh
You are now setup for bastion usage
2, https://w.amazon.com/index.php/GUAM
To reload user and group permissions on a particular host, running auth-sync.pl will pull the latest information from the user management servers. It can be run manually with the following command:
sudo /usr/multipass/bin/auth-sync.pl -a
Or, to sync an entire hostclass (50 hosts at a time), use envImprovement's RemoteCommand:
/apollo/env/envImprovement/bin/remote-command --class NAME-OF-HOSTCLASS --recurse "sudo /usr/multipass/bin/auth-sync.pl -a" --batch 50
You might get an error if you rebooted in single user mode in order to run Guam. This is due to the network not being active. You can start network support on RHEL using this command:
/sbin/service network start
or
sudo /sbin/service network start
If the above does not work, try remounting the root partition as read-write:
mount -o remount,rw /
3, this one is long
https://tt.amazon.com/0050033524
4, I run CMFV2 Manually::https://w.amazon.com/index.php/HowToManuallyRunCMF
HowToManuallyRunCMF
Having problems logging in? Check out Linux/Desktop/Login problems.
Contents
|
If you need to run cmf outside of its normal crontab run in the middle of the night, run it like so:
$ sudo /var/cfengine/bin/cmfv2 --verbose
the crontab adds a --sleep flag, which adds a random sleep of up to 12 hours to stagger the load on the cmf servers. you don't want the --sleep flag if you're running it manually.
You can run doldap to run cmf for a particular hostclass:
example:
$ sudo /opt/systems/bin/doldap --printhost --threads 20 \ --recurse --hostclass RETAIL-UK-ONLINE-GP-DUB3-CELL1-PODE \ ssh %h '/var/cfengine/bin/cmfv2 --verbose '
I just made changes to update for CMFv2. Please note that CMFv2 provides no output by default (thus the --verbose, which is optional). Also note that CMFv2 will run CMF.
If you need to run cmf manually because, say, /etc/sudoers is missing, you will need to log in as root, or reboot in single mode. To do the latter, reboot the box. Hit F9 when you're offered the "F9=BOOT MENU" option (which may not happen, in which case just wait for the big red boot menu to come up on its own. -willp). When the LILO prompt appears, type:
LILO boot: linux single
(You need to type "custom single" if the image is called "CUSTOM".) This should load the single-user mode. When the prompt appears, start up the network and start cmfv2 by running the following commands:
sh-2.05# service network restart
sh-2.05# /var/cfengine/bin/cmfv2
NOTE: This only applies to hosts running RHEL3!
If /opt/disco/cmf/bin/cmf is missing, you'll need to attempt a re-bootstrap of CMF. Execute the following as root:
sh-2.05# /opt/disco/bin/bootstrap -t cmf
If you get a message like:
Not bootstrapping on an unsupported OS (Linux-2.6c2.5-i686)
then you're trying to bootstrap Legacy CMF on a host running >RHEL3.
GUAM can also be run this way:
$ sudo /opt/disco/guam/bin/guam-update
At least, this works on Amazon Linux 2012.03.
sudo /opt/disco/bin/bootstrap -t guam
I've found that if CMF hadn't run automatically on RHEL, Kerberos also may not be setup properly for various applications.
See Kerberos_FireFox and Kerberos_Use
The key is : "dont`t let CMFv2 revise your /etc/resolve.conf" by using : chattr +a /etc/resolve.conf (disable delete, add only)
5, I met "the system administrator has disabled your account"
6, tried maually update GUAM (Global User Account Management) no effect.
7, tried revise the /etc/passwd home /bin/false to /bin/zsh... no effect....changing back to /bin/false
8, add people log in according to.... https://permissions.amazon.com/hostclass.mhtml?target=1939662
9, trying to remove the home dir overide https://permissions.amazon.com/hostclass.mhtml?action=finished&first_last=&target=3286221&entity=490355&show_login=&request=73045901
according to https://tt.amazon.com/0047470546
add root access,,,,, remember don`t overide the home dir!!!!