[
https://issues.apache.org/jira/browse/CLOUDSTACK-4184?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13733409#comment-13733409
]
Jayapal Reddy commented on CLOUDSTACK-4184:
-------------------------------------------
Reproducing Steps:
Method2:
1. Deploy vm with password enabled template.
2. Log in with the cloudstack generated password.
3. Reset the password for the vm (cloud-set-guest password received empty
string as passed)
4. Login with the cloudstack generated password.
5. To run the script on vm again when log in failed, login by ssh key.
Method2:
1. Deploy 20 parallel VMs with password enabled template. (add set -x to
cloud-set-password to see logs)
2. Observe the VM boot logs.
3. For some of the VMs the password reset failed.
4. For some VMs the password script set the password but failed to login with
password.
When I deploy parallel vms for some of the VMs the password is set properly (In
vm boot logs observed the sending save_password to router) and for some of the
vms password set failed (There is no save_pasword command send to router).
Added the below in serve_password.sh to enable debug.
set -x
exec 3>&0 4>&1 >> /var/log/test.log 2>&1
Run the cloud-set-guest-password from the VM, obseved the test.log in router.
In router:
set -x
exec 3>&0 4>&1 >> /var/log/test.log 2>&1
In guest VM: empty password received
wget -q -t 3 -T 20 -O - --header 'DomU_Request: send_my_password'
10.1.1.1:8080
+ password=
++ echo
> VM password reset works inconsistently
> --------------------------------------
>
> Key: CLOUDSTACK-4184
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-4184
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: Virtual Router
> Affects Versions: 4.2.0
> Reporter: Jayapal Reddy
> Priority: Blocker
> Fix For: 4.2.0
>
>
> 1. When password reset fails for one vm then password reset is not working
> then on.
> 2. In router the password entries are made properly.
> 3. serve password script is giving the password correctly but the vm did not
> recieved it
> Here are the logs:
> === serve_password.sh debug logs
> + PASSWD_FILE=/var/cache/cloud/passwords
> + ip=10.1.1.143
> + logger -t cloud 'serve_password called to service a request for 10.1.1.143.'
> + read input
> + '[' 'GET / HTTP/1.0' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo GET / HTTP/1.0
> + request=
> + '[' '' '!=' '' ']'
> + read input
> + '[' 'User-Agent: Wget/1.11.4 Red Hat modified' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo User-Agent: Wget/1.11.4 Red Hat modified
> + request=
> + '[' '' '!=' '' ']'
> + read input
> + '[' 'Accept: */*' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo Accept: redundant_router/arping_gateways.sh.templ
> redundant_router/backup.sh.templ redundant_router/check_bumpup.sh
> redundant_router/check_heartbeat.sh.templ
> redundant_router/checkrouter.sh.templ redundant_router/conntrackd.conf.templ
> redundant_router/disable_pubip.sh redundant_router/enable_pubip.sh.templ
> redundant_router/fault.sh.templ redundant_router/heartbeat.sh.templ
> redundant_router/keepalived.conf.templ redundant_router/master.sh.templ
> redundant_router/primary-backup.sh.templ redundant_router/services.sh
> + request=
> + '[' '' '!=' '' ']'
> + read input
> + '[' 'Host: 10.1.1.1:8080' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo Host: 10.1.1.1:8080
> + request=
> + '[' '' '!=' '' ']'
> + read input
> + '[' 'Connection: Keep-Alive' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo Connection: Keep-Alive
> + request=
> + '[' '' '!=' '' ']'
> + read input
> + '[' 'DomU_Request: send_my_password' == '' ']'
> ++ sed 's/^[ \t]*//'
> ++ cut -d: -f2
> ++ grep DomU_Request:
> ++ echo DomU_Request: send_my_password
> + request=send_my_password
> + '[' send_my_password '!=' '' ']'
> + break
> + '[' send_my_password == send_my_password ']'
> ++ get_value /var/cache/cloud/passwords 10.1.1.143
> ++ local filename=/var/cache/cloud/passwords
> ++ local keyname=10.1.1.143
> ++ cut -d= -f2
> ++ grep -i 10.1.1.143= /var/cache/cloud/passwords
> + password=bG9wrskhw
> + '[' bG9wrskhw == '' ']'
> + logger -t cloud 'serve_password sent a password to 10.1.1.143.'
> + echo bG9wrskhw
> bG9wrskhw
> + unlock_exit 0 passwdlock 1
> + releaseLockFile passwdlock 1
> + __LOCKDIR=/tmp
> + __LOCKFILE='/tmp/*-29091-passwdlock.lock'
> + __locked=1
> + '[' 1 == 1 ']'
> + rm /tmp/1375947770244011718-29091-passwdlock.lock
> + exit 0
> ^C
> root@r-36-QA:~# ps aux | grep socat
> root 29139 0.0 0.8 4788 1024 pts/0 S 07:42 0:00 socat -lf
> /var/log/cloud.log TCP4-LISTEN:8080,reuseaddr,crnl,bind=10.1.1.1
> SYSTEM:/opt/cloud/bin/serve_password.sh "$SOCAT_PEERADDR"
> root 29141 0.0 0.6 3796 804 pts/0 S+ 07:43 0:00 grep socat
> root@r-36-QA:~#
> ====== Guest vm cloud-set-guest-password script logs
> PASSWORD_SERVER_IP=10.1.1.1
> + '[' -n 10.1.1.1 ']'
> + logger -t cloud 'Found password server IP 10.1.1.1 in
> /var/lib/dhclient/dhclient-eth0.leases'
> + logger -t cloud 'Sending request to password server at 10.1.1.1'
> ++ wget -q -t 3 -T 20 -O - --header 'DomU_Request: send_my_password'
> 10.1.1.1:8080
> + password=
> ++ echo
> ++ tr -d '\r'
> + password=
> + '[' 0 -eq 0 ']'
> + logger -t cloud 'Got response from server at 10.1.1.1'
> + case $password in
> + logger -t cloud 'Password server at 10.1.1.1 did not have any password for
> the VM'
> + continue
> + for DHCP_FILE in '$DHCP_FOLDERS'
> + '[' -f '/var/lib/dhcp3/*' ']'
> + '[' 0 == 0 ']'
> + '[' 0 == 1 ']'
> + logger -t cloud 'Did not need to change password.'
> + exit 0
> [root@v3 init.d]#
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
