3e8cd1b67d
We've run into what appears to be a race with apache trying to reuse a
pooled connection to a backend when that pool connection is closing.
This leads to errors like:
[Fri Dec 07 21:44:10.752362 2018] [proxy_http:error] [pid 19073:tid 139654393218816] (20014)Internal error (specific information not available): [client 104.130.127.213:45408] AH01102: error reading status line from remote server 127.0.0.1:60999
[Fri Dec 07 21:44:10.752405 2018] [proxy:error] [pid 19073:tid 139654393218816] [client 104.130.127.213:45408] AH00898: Error reading from remote server returned by /image/v2/images/ec31a4fd-e22b-4e97-8c6c-1ef330823fc1/file
According to the internets this can be addressed (at the cost of some
performance) by setting the proxy-initial-not-pooled env var for mod
proxy. From the mod_proxy docs:
If this variable is set, no pooled connection will be reused if the client
request is the initial request on the frontend connection. This avoids the
"proxy: error reading status line from remote server" error message caused
by the race condition that the backend server closed the pooled connection
after the connection check by the proxy and before data sent by the proxy
reached the backend. It has to be kept in mind that setting this variable
downgrades performance, especially with HTTP/1.0 clients.
Closes-Bug: #1807518
Change-Id: I374deddefaa033de858b7bc15f893bf731ad7ff2
(cherry picked from commit
|
||
|---|---|---|
| data | ||
| doc | ||
| exercises | ||
| extras.d | ||
| files | ||
| gate | ||
| inc | ||
| lib | ||
| pkg | ||
| playbooks | ||
| roles | ||
| samples | ||
| tests | ||
| tools | ||
| .gitignore | ||
| .gitreview | ||
| .mailmap | ||
| .zuul.yaml | ||
| FUTURE.rst | ||
| HACKING.rst | ||
| LICENSE | ||
| MAINTAINERS.rst | ||
| Makefile | ||
| README.rst | ||
| clean.sh | ||
| exercise.sh | ||
| exerciserc | ||
| functions | ||
| functions-common | ||
| openrc | ||
| run_tests.sh | ||
| setup.cfg | ||
| setup.py | ||
| stack.sh | ||
| stackrc | ||
| tox.ini | ||
| unstack.sh | ||
README.rst
DevStack is a set of scripts and utilities to quickly deploy an OpenStack cloud from git source trees.
Goals
- To quickly build dev OpenStack environments in a clean Ubuntu or Fedora environment
- To describe working configurations of OpenStack (which code branches work together? what do config files look like for those branches?)
- To make it easier for developers to dive into OpenStack so that they can productively contribute without having to understand every part of the system at once
- To make it easy to prototype cross-project features
- To provide an environment for the OpenStack CI testing on every commit to the projects
Read more at https://docs.openstack.org/devstack/latest
IMPORTANT: Be sure to carefully read stack.sh and any other scripts you execute before you run them, as they install software and will alter your networking configuration. We strongly recommend that you run stack.sh in a clean and disposable vm when you are first getting started.
Versions
The DevStack master branch generally points to trunk versions of OpenStack components. For older, stable versions, look for branches named stable/[release] in the DevStack repo. For example, you can do the following to create a Pike OpenStack cloud:
git checkout stable/pike
./stack.shYou can also pick specific OpenStack project releases by setting the
appropriate *_BRANCH variables in the
localrc section of local.conf (look in stackrc for the default set). Usually just
before a release there will be milestone-proposed branches that need to
be tested:
GLANCE_REPO=git://git.openstack.org/openstack/glance.git
GLANCE_BRANCH=milestone-proposedStart A Dev Cloud
Installing in a dedicated disposable VM is safer than installing on your dev machine! Plus you can pick one of the supported Linux distros for your VM. To start a dev cloud run the following NOT AS ROOT (see DevStack Execution Environment below for more on user accounts):
./stack.sh
When the script finishes executing, you should be able to access OpenStack endpoints, like so:
- Horizon: http://myhost/
- Keystone: http://myhost/identity/v2.0/
We also provide an environment file that you can use to interact with your cloud via CLI:
# source openrc file to load your environment with OpenStack CLI creds
. openrc
# list instances
openstack server listDevStack Execution Environment
DevStack runs rampant over the system it runs on, installing things and uninstalling other things. Running this on a system you care about is a recipe for disappointment, or worse. Alas, we're all in the virtualization business here, so run it in a VM. And take advantage of the snapshot capabilities of your hypervisor of choice to reduce testing cycle times. You might even save enough time to write one more feature before the next feature freeze...
stack.sh needs to have root access for a lot of tasks,
but uses sudo for all of those tasks. However, it needs to
be not-root for most of its work and for all of the OpenStack services.
stack.sh specifically does not run if started as root.
DevStack will not automatically create the user, but provides a
helper script in tools/create-stack-user.sh. Run that (as
root!) or just check it out to see what DevStack's expectations are for
the account it runs under. Many people simply use their usual login (the
default 'ubuntu' login on a UEC image for example).
Customizing
DevStack can be extensively configured via the configuration file local.conf. It is likely that you will need to provide and modify this file if you want anything other than the most basic setup. Start by reading the configuration guide for details of the configuration file and the many available options.