[Zope] ZEO clients aren't duplicated

[Christopher Muldrow]

I thought the whole point of ZEO was to allow a change on one client to 
be available immediately on all clients. That's certainly the way ZEO 1 
worked--we'd add a document through a single client and it would appear 
on all clients as soon as it was added. Now we're adding documents on 
one and not seeing them on the other three. We'd have to add some way 
to "tickle" the clients every time we make a change to the site if this 
is the way ZEO 2 works. Is that right?


On Tuesday, October 7, 2003, at 07:03 PM, J Cameron Cooper wrote:

>
>> We just did an upgrade to Zope 2.6.2 and ZEO 2 (we were on 2.6.1 and  
>> ZEO 1 before). We ended up setting up clients on port 8080 this time, 
>>  rather than running them on 80 as we did before, but we're using the 
>>  same round robin Squid load balancing setup we were using previously 
>>  (from Toby Dickenson's howto here:  
>> http://www.zope.org/Members/htrd/howto/squid). What we're noticing,  
>> though, is now it looks like changes we make through our rewritten 
>> URLs  are occurring on one client and not necessarily happening 
>> immediately  on the other clients. I'm guessing this is a caching 
>> issue or something  that we didn't have to deal with with ZEO 2, but 
>> I'm not sure.
>
> A ZEO client must make a connection to the server to learn about dirty 
> objects so that it can flush its cache, since there is no way for the 
> server to contact its clients. Or so I read once, and it makes sense.
>
> Try doing something on the other client that will cause it to contact 
> the server for some reason, and see if that makes the updates visible.
>
>                --jcc
>
> -- 
> "Code generators follow the 80/20 rule. They solve most of the 
> problems, but not all of the problems. There are always features and 
> edge cases that will need hand-coding. Even if code generation could 
> build 100 percent of the application, there will still be an endless 
> supply of boring meetings about feature design."
> (http://www.devx.com/java/editorial/15511)
>
>
>