init process does not start, but I can start exhibitor from the command line about zookeeper-cookbook

[mwhooker]

It might be best to try the latest tagged version, 1.4.9, until we can troubleshoot.

thanks for the report.

[alkhatib]

Would the latest 1.4.9 be the same version that is in the chef community
cookbook?

Because that is the one I initially tried before trying the latest commit.
And I got the same problem.

Thanks
On 27 Jan 2014 16:54, "Matthew Hooker" [email protected] wrote:

It might be best to try the latest tagged version, 1.4.9https://github.com/SimpleFinance/chef-zookeeper/tree/v1.4.9until we can troubleshoot.

thanks for the report.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/51#issuecomment-33427540
.

[mwhooker]

yeah, 1.4.9 will be latest community cb. That's good to know.

Debugging now, I'll update you soon

[mwhooker]

Having trouble reproducing.

check-local-zk.py blocks the init script until it can reach zookeeper. It tries for 5 minutes. When you run exhibitor manually, it doesn't run that script, which is why it's starting okay.

Could you run that script after starting exhibitor manually and tell me what the output is?

It would also be useful to see the changes you made to the attributes.

[alkhatib]

Sure, I did this on a new machine Amazon Linux machine:
Added ec2-user to the sudoers file:

sudo visudo
+ ec2-user ALL=(ALL)      ALL

Did a yum update and installed patch which is needed by the recipe (Maybe a dependency needs to be defined somewhere?):
sudo yum update sudo yum install patch

Make the parent directory that contains my snapshot,logs/transaction directories:
mkdir ~/zookeeper

My only changes were made to default.rb and my diff of the default.rb file:

nottaway:zookeeper aalkhatib$ git diff
diff --git a/attributes/default.rb b/attributes/default.rb
index 8045329..8ff515f 100644
--- a/attributes/default.rb
+++ b/attributes/default.rb
@@ -2,8 +2,8 @@ default[:zookeeper][:version] = "3.4.5"
 default[:zookeeper][:mirror] = "http://mirrors.ibiblio.org/apache/zookeeper/zookeeper-#{default[:zookeeper][:version]}/zookeeper-#{default[:zookeeper][:version]}.tar.gz"
 default[:zookeeper][:checksum] = 'e92b634e99db0414c6642f6014506cc22eefbea42cc912b57d7d0527fb7db132'
 default[:zookeeper][:install_dir] = "/opt/zookeeper"
-default[:zookeeper][:user] = "zookeeper"
-default[:zookeeper][:group] = "zookeeper"
+default[:zookeeper][:user] = "ec2-user"
+default[:zookeeper][:group] = "ec2-user"

 default[:gradle][:version] = "1.5"
 default[:gradle][:mirror] = "http://services.gradle.org/distributions/gradle-#{default[:gradle][:version]}-bin.zip"
@@ -14,9 +14,9 @@ default[:exhibitor][:install_dir] = "/opt/exhibitor"

 default[:exhibitor][:script_dir] = '/usr/local/bin/'

-default[:exhibitor][:snapshot_dir] = "/tmp/zookeeper"
-default[:exhibitor][:transaction_dir] = "/tmp/zookeeper"
-default[:exhibitor][:log_index_dir] = "/tmp/zookeeper_log_indexes"
+default[:exhibitor][:snapshot_dir] = "/home/ec2-user/zookeeper/snapshots"
+default[:exhibitor][:transaction_dir] = "/home/ec2-user/zookeeper/txn"
+default[:exhibitor][:log_index_dir] = "/home/ec2-user/zookeeper/logs"
 default[:exhibitor][:log_to_syslog] = "1"

 # Port for the HTTP Server
@@ -24,15 +24,13 @@ default[:exhibitor][:opts][:port] = "8080"
 default[:exhibitor][:opts][:hostname] =  node[:ipaddress]
 default[:exhibitor][:opts][:defaultconfig] = "#{node[:exhibitor][:install_dir]}/exhibitor.properties"

-default[:exhibitor][:opts][:configtype] = "file"
+default[:exhibitor][:opts][:configtype] = "s3"

 default[:exhibitor][:loglevel] = "info"

 # For --configtype s3, set:
-# [:exhibitor][:s3key] = "key"
-# [:exhibitor][:s3secret] = "secret"
-# [:exhibitor][:opts][:s3config] = "bucket:config-key"
-# [:exhibitor][:opts][:s3region] = "region" # i.e. us-east-1
+default[:exhibitor][:s3key] = "MY-KEY"
+default[:exhibitor][:s3secret] = "MY-SECRET"
+default[:exhibitor][:opts][:s3config] = "MYBUCKET:KEY"
+default[:exhibitor][:opts][:s3region] = "us-east-1" # i.e. us-east-1

 # For --contiftype file
 default[:exhibitor][:opts][:fsconfigdir] = "/tmp"

Then I start the bootstrap:

knife cookbook upload zookeeper
Uploading zookeeper      [1.4.9]
Uploaded 1 cookbook.

knife bootstrap x.x.x.x --ssh-user ec2-user --run-list "recipe[zookeeper]" --sudo

[alkhatib]

So it actually hung at:

  * remote_file[/var/chef/cache/zookeeper-3.4.5.tar.gz] action create

I checked the machine and no chef process was running, so I just restarted it, and will update when it is finished.

[alkhatib]

Ok, so the bootstrap hung again at the point where it was starting the process and waitiing for the check-local-zk.py script to return.

In /var/log/messages:

Jan 28 16:02:48 init: exhibitor main process (31063) terminated with status 1
Jan 28 16:02:48 init: exhibitor main process ended, respawning
Jan 28 16:07:49 init: exhibitor post-start process (31064) terminated with status 1
Jan 28 16:07:49 init: exhibitor main process (31077) terminated with status 1

Then when I started the zookeeper instance manualy by copy/pasting the command in /etc/init/exhibitor.conf:

sudo -u $USER bash -c "java -Dlog4j.configuration=file:///opt/exhibitor/log4j.properties -jar /opt/exhibitor/1.5.0.jar \
    --configtype s3 --defaultconfig /opt/exhibitor/exhibitor.properties --hostname x.x.x.x --port 8080 --s3config bucket:name --s3credentials /opt/exhibitor/exhibitor.s3.properties --s3region us-east-1 | logger -t zookeeper"

That worked and I was able to connect to zookeeper, and the s3 configuration was updated, and that machine joined my ensemble.
and the check-local-zookeeper.py exits with a 0 status

[ec2-user@machine ~]$ /usr/local/bin/check-local-zk.py
[ec2-user@machine ~]$ echo $?
0

Killing the exhibitor/zookeeper I started and trying to start it as a service:

sudo start -v exhibitor

/var/log/messages:
Jan 28 16:26:32 init: exhibitor main process (31611) terminated with status 1
Jan 28 16:26:32 init: exhibitor main process ended, respawning

and the script:
python /usr/local/bin/check-local-zk.py
<urlopen error [Errno 111] Connection refused>
<urlopen error [Errno 111] Connection refused>
<urlopen error [Errno 111] Connection refused>

If you need anymore information, or would like me to do anything let me know.

[alkhatib]

Any other information I can provide to help debug this issue?

[mwhooker]

let me try with an Amazon Linux image

[mwhooker]

dunno why this didn't click, but it's probably because this cookbook supports upstart, which is an ubuntu thing. You might try applying the patches in #19

I'll see if I can add support for runit

[mwhooker]

closing. Please track in #54