On 06/11/2004 Andreas Jung wrote:
--On Freitag, 5. November 2004 John Poltorak wrote:
In the process of migrating from 2.6.2 to 2.7.3 can I run both versions temporarily? If so how?
By installing Zope in different locations and running Zope on different ports. Just a question of configuration...
i'm doing exactly this right now: we run a zope 2.6.4 debian/testing installation with several instances and python2.2. besides that we run a zope 2.7.2 debian installation with python2.3. now we want to add a third installation, zope 2.6.4 from source installation with python2.1, as we encounter some strange problems that we aren't able to locate, and we just need to test whether it has to do with the debian packages and/or python2.2. zope 2.6.4 recomments python2.1, and i've no glue whether the debian packages run it with python2.2, as python2.1 is in the debian archive as well. my question for that is: when we install the third zope installation, is there any problem with just linking Products from the python2.2 zope 2.6.4 Products folder to the new one? i could imagine that the different python versions generate troubles, as the scripts are compiled with 2.2? but maybe the work we're doing here is senceless, and this bug is already well known to you all, so i'll try to formulate a brief description of it. sometimes, not really localizable when, but apparently when the zope instance is under some load, it silently disappears. the system processes are still around, but i don't know whether _all_ of them are still there. the zope instance HTTP and FTP servers are completely dead. zopectl restart doesn't work, as it doesn't kill the old processes, only manually killing them and starting the instance afterwards is an option. we weren't able to reproduces this bug anywhere else yet. not with zope 2.6.4 server under windows nt, and even not with another debian/testing zope 2.6.4 installation, with exactly the same products. could you imagine that this problems concerns python2.2 usage? bye jonas