[ZODB-Dev] ZEO on NT upgrade hints?

[Dunigan, Craig]

Thanks very much for your response.  Comments and answers inline below.

> >>>>> "CD" == Dunigan, Craig <craig.dunigan@esker.com> writes:
> 
>   CD> I just tried to upgrade ZEO from 0.4 to 1.0b3, and ran into a
>   CD> few problems.  First, it's clear that this version of ZEO wants
>   CD> v2.3.2 of Zope available (I'm using 2.3.0), because I got many
>   CD> "cannot import up-to-date cPickle" errors until I tried it with
>   CD> 2.3.2.
> 
> Sorry about that.  I'm not sure what versions of cPickle and asyncore
> ship with what versions of Zope.  Perhaps Jim can chime in with an
> explanation of what the minimum requirements for ZEO 1.0 are, in which
> case I'll update the README file.

Thanks!
 
>   CD> D:\Zope\bin\python D:\Zope\lib\python\ZEO\start.py -p 8800
> 
>   CD> I have a plain vanilla ZODB FileStorage living at var/Data.fs.
>   CD> There is no Zope process running on the ZEO server, just ZEO.
>   CD> Running the above batch file now returns to the command prompt
>   CD> (which it shouldn't) without any errors, but without ZEO
>   CD> running, either (no python processes in Task Manager).  I'm
>   CD> really not clear on how to proceed from here.
> 
> The zLOG may contain more useful information.  I'm not sure how to
> invoke it on NT, but I'll describe how I use it on Unix and hope you
> can translate to your native tongue.  zLOG depends on two environment
> variables, STUPID_LOG_FILE and STUPID_LOG_SEVERITY.  (The names are
> stupid, I know.)  If you define STUPID_LOG_FILE, then zLOG output will
> be written to a file with that name; if you use set the value to '',
> the output will be sent to stderr.  The STUPID_LOG_SEVERITY controls
> which error messages are ignored and which are logged.  The severity
> levels range from -300 to 300 in increments of 100.  The default value
> is 0.  I'd recommend setting it to -300 to get all the output.

I'll sure do this.  For the record, I'm more of a Linux person myself.  Our
IS people wanted the ZEO server on NT so they could fit it into their
existing backup scheme, and so that we wouldn't hit the ext2fs 2GB file size
limit.  A little ambitious, perhaps (we're at about 200MB after four months
use), but the word "limit" makes them nervous.  

> If you can record the zLOG output for a failed attempt to run
> start.py and send it to me, I may be able to help debug it.
> 
> Jeremy

Jeremy, I am overflowing with gratitude, and that's not sarcastic!  I'll get
back to you shortly.

Craig Dunigan
Web Programmer
Esker Software - Extending the Reach of Information
mailto:craig.dunigan@esker.com
Ph. 608.273.6000
Fax 608.273.8227
http://www.esker.com