OpenVPN Troubleshooting

Recently I tried installing and setting up OpenVPN on my home server. I ran into a problem I had had before. Both the installation of the .deb package and the init script would hang.

I found the –verbose switch to the /usr/local/openvpn_as/bin/ovpn-init command. it would always hang right after “Generating init scripts auto command”. As generating the init scripts should be instantaneous, this was frustrating. At one point in the script is calls /usr/local/openvpn_as/bin/_ovpn-init and attempts to list the iptables config with iptable –list and put it into a variable. My iptables list is LARGE and uses the hostnames of several IPs I want blocked. (Think asian and Russian sites.) That iptables command it uses does DNS on every hostname. It should use (and I changed the script) to use iptables –list -n which skips DNS.

Now the ovpn-init script runs and doesn’t hang.

Zabbix and Percona

I am staging a Zabbix v3 install and needed to monitor MySQL with Percona. I used the Zabbix Template from this page, and ran into errors:

Invalid tag “/zabbix_export/templates/template(1)/items/item(1)”: the tag “snmp_community” is missing.

Among others.

It looks like the template from Percona is a bit outdated for Zabbix v3. I spent a few hours fixing it. Here it is, fixed.


Continue reading Zabbix and Percona

iOS 8.0 and excessive ‘Documents & Sync’ over LTE

“Verizon Msg: You’ve used about 50% of your 3GB data plan…”

I received this warning only 2 days before my data plan started over. I thought nothing about it at the time. Then, a few days later, after the start of the billing cycle, I got:

“Verizon Msg: You’ve used about 50% of your 3GB data plan…”

That didn’t seem right. A Gig and a half in about 20 days?! I hadn’t used any streaming service lately – especially over LTE. But, I was able to add another Gig to my plan for free, so I thought nothing of it. But, a week later, after the start of another billing cycle:

“Verizon Msg: You’ve used about 75% of your 4GB data plan”

No! This will not stand! I delved into the Cellular settings in the Settings app and found that ‘Documents and Sync’ was using way too much data – over LTE even when I was using Wifi! I called Apple and started a ticket. But, that takes time for them to come up with a solution. So, I started to document and troubleshoot. I disabled iCloud settings and found that nothing short of disabling iCloud on the device would still cause ‘data bleeding’ as I called it.

I could see it use a Meg a minute!
I could see it use a Meg a minute!

I surmised that something was wrong with some basic sync between my other iCloud-enabled devices – even over LTE – with them on the same Wifi network! At some point, I came to the and saw the iCloud Keychain.

Screenshot 2015-04-10 12.39.58

I emptied it out. I wasn’t able to backup anything in it. After about 30 minutes to make sure it wasn’t a convenient pause, the ‘data bleeding’ was stopped!

Here’s the interesting/scary thing: I had disabled Keychain sync on the iPhone and it still tried to sync!

I hope that was a bug and not how it really works!

Now, after 4 days after my data plan has restarted, I am seeing better cellular usage – and battery usage!

This is after 4 days!
This is after 4 days!


Also fixed after this, I usually had old Wifi networks in my Preferred Networks list in the System Preferences app. They were now gone!


A Quick Note on GitLab 7.4 and Remote Storage Backups

So, 7.4 came out today and among other things, includes sending backups to remote storage like  Amazon S3 storage.

Here’s a quick tip or two about getting it setup.

Example from my gitlab.yml:

 ## Backup settings
  path: "tmp/backups" # Relative paths are relative to Rails.root (default: tmp/backups/)
  keep_time: 604800 # default: 0 (forever) (in seconds)
  # Fog storage connection settings, see .
      provider: AWS
      region: us-east-1
      aws_access_key_id: _Access-ID-ID_
      aws_secret_access_key: '_Access_KEY_'
    # The remote 'directory' to store your backups. For S3, this would be the bucket name.
    remote_directory: 'unique_backup_bucket'

I kept hitting 2 walls when getting things working.

First, the region. In the docs they use a EU region. I couldn’t recall the region for the default USA region. Here are the S3 regions.

Second, was whitespace. .yml files are picky.

Settingslogic::MissingSetting: Missing setting 'remote_directory' in 'upload' section in 'backup' section in /home/git/gitlab/config/gitlab.yml
/home/git/gitlab/vendor/bundle/ruby/2.1.0/gems/settingslogic-2.0.9/lib/settingslogic.rb:189:in `missing_key'
/home/git/gitlab/vendor/bundle/ruby/2.1.0/gems/settingslogic-2.0.9/lib/settingslogic.rb:117:in `method_missing'
/home/git/gitlab/lib/backup/manager.rb:33:in `upload'
/home/git/gitlab/lib/backup/manager.rb:29:in `pack'
/home/git/gitlab/lib/tasks/gitlab/backup.rake:15:in `block (3 levels) in '
Tasks: TOP => gitlab:backup:create
(See full trace by running task with --trace)

This means you have extra whitespace before the remote_directory stanza. It should align with the connection stanza.

A Quick Fix for a minor GitLab GUI Error

So I recently encountered a weird error in the Admin Dashboard of GitLab:

No GitLab-Shell Version

It wasn’t a gitlab-shell error, the bin/check script ran fine. Git and Ruby at accepted versions from the rake gitlab:check command but it did show an error:

GitLab Shell version &gt;= 1.7.9 ? ... <span style="color: #ff0000;">FAIL. Please update gitlab-shell to 1.7.9 from Unknown</span>

The fix is simple, a wrong path in gitlab.yml:

187   ## GitLab Shell settings
188     gitlab_shell:
189     path: /wrong/git/gitlab-shell/

(Some of the machine I administer use a different /home path.)

So, simply fix the path, restart gitlab and things will work.

Screenshot 2014-01-30 12.34.33


Clef two-factor authentication
%d bloggers like this: