[Erp5-dev] Moved Data.fs/parts of it - problems with site

[Łukasz Nowak]

Hello,

When I moved Data.fs or imported/exported by zexp parts of another site
to my development site I've got such problems, while accessing imported
modules:

===
This object is broken because the ERP5Type product that created it is
no longer installed or is installed incorrectly. Please contact the
product maintainer for assistance.

Note that the data associated with this product has not been lost, and
will be accessible again if the product is reinstalled.
===

But this instance, when site is created by hand is working correctly,
without any problems. There is not trail in logs or zopectl fg about
any problems.

Sometimes few restarts of Zope helps, but problem is re-appearing.

Removing all but Data.fs do not helps.

Removing all .pyc do not helps. (Well - helps for a while...)

Packing Data.fs do not helps.

This problem is sometimes appears just after starting zope or after
few minutes/hours working in zope. And sometimes disappears without any


I heard, that Data.fs is quite compatible - it might be moved from one
architecture to another, from one to another version. In my case I'm
moving Data.fs (or its parts) from 32-bit to 32-bit, from:

(Zope 2.7.8-final, python 2.4.1, linux2)
2.4.1 (#2, Aug 25 2005, 18:20:57) [GCC 4.0.1 (4.0.1-2mdk for Mandriva
Linux release 2006.0)]

to:

(Zope 2.7.8-final, python 2.4.4, linux2)
2.4.4 (#2, Apr 5 2007, 20:11:18) [GCC 4.1.2 20061115 (prerelease)
(Debian 4.1.1-21)]

Has anyone spotted same problem? Any suggestions, what might be wrong?

I've got ERP5Type product, so cited above text is quite cryptic for me.

Regards,
Luke

[Jérome Perrin]

Post by Łukasz Nowak
Hello,
When I moved Data.fs or imported/exported by zexp parts of another site
to my development site I've got such problems, while accessing imported
===
This object is broken because the ERP5Type product that created it is
no longer installed or is installed incorrectly. Please contact the
product maintainer for assistance.
Note that the data associated with this product has not been lost, and
will be accessible again if the product is reinstalled.
===
But this instance, when site is created by hand is working correctly,
without any problems. There is not trail in logs or zopectl fg about
any problems.

Hello,

Do you have local documents in $INSTANCE_HOME/Documents that you might
have forgot to copy ? (anyway this might at least log something like
'ZODB could not import class')

Post by Łukasz Nowak
Sometimes few restarts of Zope helps, but problem is re-appearing.

Then, this might be related to TimerService. RPM for mandriva is quite
old, we suspect svn version to be buggy. Can you try with r11930 of
timerserver ? (this requires installing the python package with setup.py
install)

https://svn.erp5.org/repos/public/erp5/trunk/products/TimerService/timerserver/

J?rome

[Łukasz Nowak]

Hello,

On 2007-11-28, 10:01:30

Post by Łukasz Nowak

Post by Łukasz Nowak
But this instance, when site is created by hand is working
correctly, without any problems. There is not trail in logs or
zopectl fg about any problems.

Hello,
Do you have local documents in $INSTANCE_HOME/Documents that you
might have forgot to copy ? (anyway this might at least log something
like 'ZODB could not import class')

None missing files in Documents. I've attached my start log.

Post by Łukasz Nowak

Post by Łukasz Nowak
Sometimes few restarts of Zope helps, but problem is re-appearing.

Then, this might be related to TimerService. RPM for mandriva is
quite old, we suspect svn version to be buggy. Can you try with
r11930 of timerserver ? (this requires installing the python package
with setup.py install)
https://svn.erp5.org/repos/public/erp5/trunk/products/TimerService/timerserver/

Ok. I loaded this from revision 11930, I'll check if it helps. This is
diff of what I had (from packages from deb4erp5 project, prepared by
me):

--- timerserver-MINE/TimerServer.py 2007-10-11 13:30:05.000000000
+0200 +++ timerserver/TimerServer.py 2007-11-29 11:50:56.000000000
+0100 @@ -6,7 +6,7 @@
import traceback

import thread
-import sys, os, errno, time, socket
+import sys, os, errno, time, socket, urllib
from StringIO import StringIO
from zLOG import LOG, INFO

@@ -55,15 +55,12 @@
# To be very sure, try to connect to the HTTPServer
# and only start after we are able to connect
while 1:
- s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
- s.settimeout(5)
+ time.sleep(5)
try:
- s.connect((ip, port))
- except (socket.error, socket.timeout):
- s.close()
+ urllib.urlopen('http://%s:%s' %(ip, port))
+ except IOError:
continue
break
- s.close()

module = self.module
interval = self.interval


Regards,
Luke

[Łukasz Nowak]

Hello,

On 2007-11-29, 11:55:10
?ukasz Nowak <lukasz.nowak at ventis.com.pl> wrote:

(...)

Post by Łukasz Nowak
Ok. I loaded this from revision 11930, I'll check if it helps. This is
diff of what I had (from packages from deb4erp5 project, prepared by

(...)

After few days working with your suggestion this problem no more
appeared. So I updated packages on code-google.

Thanks,
Luke