Yes the culprit were the git fetch operations. When a new review request
form is created, reviewboard internally fetches all git repos and that
taking over a minute to complete. With Apache 2.4 the default timeout value
is 60s as opposed to apache 2.2 (300s) which is why it started failing
Hi
My reviewboard version 1.5.1
scmtools/git.py contains the git fetch lines i am seeing (debug=True is
enabled in settings):
class GitClient:
def __init__(self, path):
self.path = path
# Adding a skinit before git fetch so that kerberos credentials
# are acquired
Hi,
Glad you found the cause, but I have to point out that this isn't out code
(it's an in-house modification at your company), and 1.5.1 is ancient and
no longer supported. We're not really going to be able to offer much help
with this server in the future, but if you're looking to upgrade to
When I click on "new review request" I see following logs in reviewboard:
...
[Mon May 16 01:17:46.025209 2016] [wsgi:error] [pid 19078] Before executing
the git fetch statement:
[Mon May 16 01:17:46.025234 2016] [wsgi:error] [pid 19078] The path is:
I get this error in reviewboard logs:
[Mon May 16 06:29:59.527802 2016] [wsgi:error] [pid 4995] [client
10.79.176.142:32246] Timeout when reading response headers from daemon
process 'sdlctool': /local/apps/reviewboard-prod/conf/apache_wsgi.conf,
referer:
Hi,
We don't do any git fetches, and we don't output that log statement (or
output to the Apache log file). Are there any modifications to your Review
Board source code that might be doing this?
You'll need to figure out what's doing that git fetch, as that definitely
sounds like it's the
Thanks for you quick reply.
What about the "Timeout when reading response headers from daemon process"
error. If I want to increase the timeout value as workaround for quick fix
will it help? If so which timeout value should be configured here?
On Monday, May 16, 2016 at 4:29:21 PM UTC+5:30,