ckrintz / appscale

Automatically exported from code.google.com/p/appscale
0 stars 0 forks source link

Voldemort doesn't work #197

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
In our launchpad main branch, Voldemort throws up the following exception:

Traceback (most recent call last):
  File "/root/appscale/AppDB/soap_server.py", line 1090, in <module>
    db = appscale_datastore.DatastoreFactory.getDatastore(datastore_type)
  File "/root/appscale/AppDB/appscale_datastore.py", line 38, in getDatastore
    d_mod = imp.load_source(d_name, mod_path)
  File "/root/appscale/AppDB/voldemort/py_voldemort.py", line 17, in <module>
    from voldemort_client.voldemort import StoreClient
  File "/root/appscale/AppDB/voldemort/voldemort_client/voldemort.py", line
10, in <module>
    import voldemort_client_pb2 as protocol
  File
"/root/appscale/AppDB/voldemort/voldemort_client/voldemort_client_pb2.py",
line 3, in <module>
    from google.protobuf import descriptor
ImportError: No module named protobuf

This appears to be due to not having installed protocol buffers installed.
Will examine how to change the scratch install accordingly.

Original issue reported on code.google.com by shattere...@gmail.com on 22 Apr 2010 at 4:16

GoogleCodeExporter commented 9 years ago
Fixed. Ran the protocol buffer install (installprotobuf_old) in
appscale_install_functions.sh and added the protocol buffer egg to the 
pythonpath:

  ENV['PYTHONPATH']="/usr/local/lib/python2.6/dist-packages/protobuf-2.3.0-py2.6.egg"
  Djinn.log_debug(`start-stop-daemon --start --background --exec
#{voldemort_loc}/bin/voldemort-server.sh -- #{voldemort_loc}/config/appscale`)

Yoshi, can you change the install to run the installprotobuf_old function upon
install of Voldemort? We also need to know what to change so that I can do this
change next time around.

Original comment by shattere...@gmail.com on 22 Apr 2010 at 4:54

GoogleCodeExporter commented 9 years ago
I think we should use debian package as much as we can, because the own binary 
might
breaks dependency of the other packages.
I think you use appscale_build.sh, so could you re-execute "appscale_build.sh" 
or
execute "appscale_install_jaunty.sh voldemort" ?

Original comment by yoshi...@gmail.com on 22 Apr 2010 at 6:06

GoogleCodeExporter commented 9 years ago
No, since the current code in the stable branch isn't tested on Euca and isn't
actually stable, I don't want to risk destroying my working images. Since I 
can't
commit code to the stable branch for the same reason, doing so will also cause 
me to
lose the work I've been doing since then.

Can you build a new image and test this for me? This fixes the problem both you 
and
Jonathan are having, so please find a way to work this solution into the 
packaging
system or suggest an alternative.

Original comment by shattere...@gmail.com on 22 Apr 2010 at 6:49

GoogleCodeExporter commented 9 years ago
I confirmed unstable branch rev 596 (same as stable branch rev 471) is working
correctly in single node and 4 nodes KVM.

Original comment by yoshi...@gmail.com on 6 May 2010 at 12:28