CentOS 7: Want a functional loggerhead

Fri Sep 11 22:12:30 UTC 2015

> -----Original Message-----
> From: Richard Wilbur [mailto:richard.wilbur at gmail.com]
> Sent: Friday, September 11, 2015 3:02 PM
>
> Sorry for the misunderstanding on my part.  Thank you, Kevin, for
> explaining the situation in more detail.  It looks like I will need to
> grab a copy of loggerhead to beat on myself.

I realize "double posting" is taboo ... hopefully this doesn't count as a
double-post.  I found a prior CentOS question on Launchpad, so for reference,
I posted the following there (before getting a response here) since it might
help others when it is finally resolved:

  https://answers.launchpad.net/loggerhead/+question/271275

I had continued to troubleshoot after posting to the list, so it added some
information that I didn't know when I posted to the list:

When I tried to use the deprecated `start-loggerhead` and `stop loggerhead`
Scripts on CentOS 7, loggerhead still doesn't work serving branches, but
I don't get a 404 error!  I get an almost blank browser screen with only
"loggerhead branches" in the top-right corner.  I realize I don't know
that my start-loggerhead environment is setup correctly as I have not
replicated it anywhere else, so here I'll show a less edited sequence
of how I tested loggerhead by mirroring what I did in CentOS 7 on my
CentOS 6 box.

What follows demonstrates what I am trying to do on CentOS 7, but everything
below is done on CentOS 6 (where loggerhead already works) to prove the
sanity of my testing procedure on CentOS 7 and to help others
potentially replicate my CentOS 7 environment:

Remove CentOS/EPEL-supplied packages not supplied by CentOS 7/EPEL 7
$ yum remove loggerhead python-simpletal

Create a local home for python site-packages that does not conflict with the
distro site-packages and so I don't have to `python setup.py install` with
sudo or root.

$ sudo mkdir /home/app
$ sudo chown krb_adm.users /home/app
$ chmod 2750 /home/app
$ mkdir ~/pkg/launchpad.net
$ cd ~/pkg/launchpad.net
$ wget 
https://launchpad.net/loggerhead/1.18/1.18.1/+download/loggerhead-1.18.1.tar.gz
$ tar -xzf loggerhead-1.18.1.tar.gz
$ cd loggerhead-1.18.1
$ python setup.py install --prefix=/home/app
$ mkdir ~/pkg/owlfish.com
$ cd ~/pkg/owlfish.com
$ wget http://www.owlfish.com/software/simpleTAL/downloads/SimpleTAL-4.3.tar.gz
$ tar -xzf SimpleTAL-4.3.tar.gz
$ cd SimpleTAL-4.3
$ python setup.py install --prefix=/home/app

Test loggerhead in this configuration:

$ export PYTHONPATH=/home/app/lib/python2.6/site-packages
$ PATH=$PATH:/home/app/bin
$ cd ~/bzr
$ serve-branches blah
DEBUG:paste.httpserver.ThreadPool:Started new worker -1220445328: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1232077968: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1244660880: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1257243792: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1267733648: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1278223504: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1288713360: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1299203216: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1309693072: Initial worker 
pool
DEBUG:paste.httpserver.ThreadPool:Started new worker -1320182928: Initial worker 
pool
serving on 0.0.0.0:8080 view at http://127.0.0.1:8080
DEBUG:paste.httpserver.ThreadPool:Added task (0 tasks queued)
INFO:loggerhead:127.0.0.1 - - [11/Sep/2015:16:21:01 -0500] "GET / HTTP/1.0" 
301 - "-" "Lynx/2.8.6rel.5 libwww-FM/2.14 SSL-MM/1.4.1 OpenSSL/1.0.0-fips"
DEBUG:paste.httpserver.ThreadPool:Added task (0 tasks queued)
INFO:loggerhead.blah:built revision graph cache: 0.012253046035766602 secs
INFO:loggerhead.blah:Getting information for ChangeLogUI: 0.25657796859741211 
secs
INFO:loggerhead:127.0.0.1 - - [11/Sep/2015:16:21:03 -0500] "GET /changes 
HTTP/1.0" 200 - "-" "Lynx/2.8.6rel.5 libwww-FM/2.14 SSL-MM/1.4.1 
OpenSSL/1.0.0-fips"
INFO:simpleTAL.XMLTemplateCompiler:Recieved skipped entity: nbsp
INFO:simpleTAL.XMLTemplateCompiler:Recieved skipped entity: laquo
INFO:simpleTAL.XMLTemplateCompiler:Recieved skipped entity: raquo
INFO:loggerhead.blah:Rendering ChangeLogUI: 0.2234799861907959 secs, 8195 bytes

In another terminal, see if the branches are published:

$ lynx http://127.0.0.1:8080

Yes, they are.  The serve-branches log above shows that it processes the blah
checkout.  The browser had the revisions listed.

Next, check the start-loggerhead, stop-loggerhead scripts just in case they
might work better on CentOS 7.  Not sure how to set it up properly, but
winging it:

$ mkdir loggerhead-files
$ cp -a /home/apps/bin/st*-loggerhead .
$ sudo vim /etc/loggerhead.conf

Create /etc/loggerhead.conf containing:

title = 'branches in loggerhead'
log.roll = 'weekly'

[debug]
    name = 'blah'
    cachepath = '/home/krb_adm/bzr/loggerhead-files'

    # if an auto_publish_folder is given, any bazaar branches discovered in
    # that folder will be published.  using each branch's 'branch.conf' to
    # provide optional descriptions and public urls is a nice way to avoid
    # having to do much configuration here.
    auto_publish_folder = '/home/krb_adm/bzr/blah'

    # if all of the branches under an auto_publish_folder are exported to
    # the same (base) public url, you can specify that url prefix here, and
    # each discovered branch's public url will be generated.
    url_prefix = 'http://127.0.0.1:8080'

I suppose that the above is NOT an appropriate loggerhead.conf setup as it
does not seem to work.

$ cd ~/bzr
$ ./start-loggerhead

WARNING!!! This script has been deprecated by serve-branches, and will be 
removed in the next release. Please migrate to serve-branches and report any 
missing features.

Launching loggerhead into the background.
PID file: /home/admin/krb_adm/bzr/loggerhead.pid

See who owns the loggerhead process:

$ ps -ef | grep `cat loggerhead.pid`
krb_adm   9217     1  1 17:00 ?        00:00:00 /usr/bin/python 
./start-loggerhead

$ lynx http://127.0.0.1:8080

This shows the same results as on CentOS 7 - a blank browser screen that
has `loggerhead branches` in the upper right corner.  So my info in the
launchpad question needs updating.  start-loggerhead results in a logs
directory being made in the CWD.

$ ls -lR logs
logs:
total 0
-rw-rw----. 1 krb_adm admin 0 Sep 11 17:03 access.log
-rw-rw----. 1 krb_adm admin 0 Sep 11 17:03 debug.log
-rw-rw----. 1 krb_adm admin 0 Sep 11 17:03 error.log

I guess that means I borked up the start-loggerhead setup - not that
that should be a distraction as it says its deprecated anyway.

$ ./stop-loggerhead

Shutting down previous server @ pid 8923.

I am able to do digging with guidance.  I've got a lot of system and
scripting XP, but not a lot of python or Bazaar specific XP.  I'm able
to run systems in VMs.  This CentOS 7 is virtualized.  I could provide
an ESXi OVF... or perhaps convert it to an open format VM image.  I
have a detailed log of what was done to stand the image up.

I'm open to continuing in either forum (list or loggerhead site on
launchpad).

Kevin R. Bulgrien