Pydra working with Ubuntu 9.10

Hello all,

I've managed to install and run pydra on my  lenovo X60s notebook
running Ubuntu 9.10. Here's my own recipe:

1. I installed most of the required packages by aptitude:

sudo aptitude install python-django python-twisted mysql-server
gnutls-bin phpmyadmin

2. I used phpmyadmin to create mysql user/db for pydra. It's fairly
easy. Just point your browser to http://localhost/phpmyadmin and from
Priviledges -> Add a new user (pydra)
Of course do't forget to create a table same as user

3. I build pydra from Git repo and just run
python setup.py install

4. Starting up the db and generating the certificates can be done as
usual from pydra/scripts:
pydra_manage syncdb
pydra_web_manage syncdb
generate_cert.sh

For simplicity I didn't put any info in last step except the no. days
for certificate expiry (I put 360)
lastly I copied the generated certificate files ca-cert.pem &
ca-key.pem to /var/lib/pydra.

5. To test everything open 5 shells and run them separately

tail -f /var/log/pydra/master.log

sudo scripts/pydra_master

sudo pydra_web_manage runserver

tail -f /var/log/pydra/node.log

sudo pydra_node

6. Now you can follow the instructions from pydra website
http://pydra-project.osuosl.org/wiki/document...

and play at http://localhost:8000/

7. Also don't forget your passwords for web/database server that you
put during installing them. For simplicity they can be kept as
pydra/pydra

That's all for today. But I have noticed a minor warning in master log:

[WARNING]  Failed to load tasks from: mapreduce (mapreduce.py) -
cannot import name SQLTableSlicer

Any clue for this ?

Thansk and best regards,
Faruque

??????????????????????????????????

Theres a bug in the MapReduce task and it won't work.  That warning is
just reporting the error.  You can just ignore it (unless you want to
use mapreduce)

-PeterM O Faruque Sarker wrote:
> That's all for today. But I have noticed a minor warning in master log:
>
> [WARNING]  Failed to load tasks from: mapreduce (mapreduce.py) -
> cannot import name SQLTableSlicer
>
> Any clue for this ?
>
> Thansk and best regards,
> Faruque
>
> ??????????????????????????????????
>

It seems to be a warning produced when a node synchronizes its code
with the master. The node has to import modules defined in the code
that is transferred from the master to the node. The code
(mapreduce.py) has an unresolved dependence problem, so Pydra
complains about that.

However, I noticed that the latest mapreduce code in git repo did not
import SQLTableSlicer. Is that bug already fixed?On Sat, Mar 27, 2010 at 12:33 AM, Peter Krenesky  wrote:
> Theres a bug in the MapReduce task and it won't work. ?That warning is
> just reporting the error. ?You can just ignore it (unless you want to
> use mapreduce)
>
> -Peter
>
> M O Faruque Sarker wrote:
>> That's all for today. But I have noticed a minor warning in master log:
>>
>> [WARNING] ?Failed to load tasks from: mapreduce (mapreduce.py) -
>> cannot import name SQLTableSlicer
>>
>> Any clue for this ?
>>
>> Thansk and best regards,
>> Faruque
>>
>> ??????????????????????????????????
>>
>
> _______________________________________________
> Pydra mailing list
> Pydra at osuosl.org
> http://lists.osuosl.org/mailman/listinfo/pydr...
>

No, it has not been fixed. 

We've been working to integrate datasource usage into the worker.  A
good portion of the datasource logic was rewritten in the process. 
ParallelTask is being worked on right now, since it is simpler than
MapReduce. Afterwards MapReduce will be updated as well.Yin QIU wrote:
> It seems to be a warning produced when a node synchronizes its code
> with the master. The node has to import modules defined in the code
> that is transferred from the master to the node. The code
> (mapreduce.py) has an unresolved dependence problem, so Pydra
> complains about that.
>
> However, I noticed that the latest mapreduce code in git repo did not
> import SQLTableSlicer. Is that bug already fixed?
>
> On Sat, Mar 27, 2010 at 12:33 AM, Peter Krenesky  wrote:
>   
>> Theres a bug in the MapReduce task and it won't work.  That warning is
>> just reporting the error.  You can just ignore it (unless you want to
>> use mapreduce)
>>
>> -Peter
>>
>> M O Faruque Sarker wrote:
>>     
>>> That's all for today. But I have noticed a minor warning in master log:
>>>
>>> [WARNING]  Failed to load tasks from: mapreduce (mapreduce.py) -
>>> cannot import name SQLTableSlicer
>>>
>>> Any clue for this ?
>>>
>>> Thansk and best regards,
>>> Faruque
>>>
>>> ??????????????????????????????????
>>>
>>>       
>> _______________________________________________
>> Pydra mailing list
>> Pydra at osuosl.org
>> http://lists.osuosl.org/mailman/listinfo/pydr...
>>
>>     
>
>
>
>