git-svn-id: https://svn.erp5.org/repos/neo/trunk@2390 71dcc9de-d417-0410-9af5-da40c76e7ee4

Was enabled accidentally in r2371.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2389 71dcc9de-d417-0410-9af5-da40c76e7ee4

But make it optional, as some tests might want to stop everything
except clients (stopping single processes is already possible
separately).

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2388 71dcc9de-d417-0410-9af5-da40c76e7ee4

Call it in client, right after starting and stopping poll thread.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2387 71dcc9de-d417-0410-9af5-da40c76e7ee4

Fixes thread flood in ZODB tests (at least).

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2386 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2385 71dcc9de-d417-0410-9af5-da40c76e7ee4

It usually happens when tests are stuck, and we want to know where it
was stuck.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2384 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2383 71dcc9de-d417-0410-9af5-da40c76e7ee4

This is because verification handler doesn't implement responses to
questions sent by a storage node after it receives the order to start.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2382 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2381 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2380 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2379 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2378 71dcc9de-d417-0410-9af5-da40c76e7ee4

This makes it easier to control logging in runner (otherwise, would
go to stderr) and functional tests (directed to its own file).

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2377 71dcc9de-d417-0410-9af5-da40c76e7ee4

This was broken by r2371: it is good not to import it at file top-level,
but it must still be provided.
Also, fix 'No handler could be found for logger "NEO"' error messages,
which where hiding log messages, by using setupLog.
Also, manually set propagate to 0 so NEO messages are not output by root
logger too.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2376 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2375 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2374 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2373 71dcc9de-d417-0410-9af5-da40c76e7ee4

This is most efficient (kernel knows what's available) & cleaner.

Of course, this is not perfect either. Still, it solved many "port already
in use" happening for me with previous implementation, for some reason.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2372 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2371 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2370 71dcc9de-d417-0410-9af5-da40c76e7ee4

There are 2 reasons:
- don't mess with default logger, NEO is not an application but a module
- basicConfig only works once, and we might want to override a previous
  setup

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2369 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2368 71dcc9de-d417-0410-9af5-da40c76e7ee4

Also, avoid looking up replicator multiple times.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2367 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2366 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2365 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2364 71dcc9de-d417-0410-9af5-da40c76e7ee4

Also, fix associated unit test, which did not detect this error because
tested storage was not assigned any partition.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2363 71dcc9de-d417-0410-9af5-da40c76e7ee4

This way, when called from "ask", displayed timestamp will be closer to
packet's timeout base timestamp.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2362 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2361 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2360 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2359 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2358 71dcc9de-d417-0410-9af5-da40c76e7ee4

This solve a race condition where the client's thread poll triggers
connection(Lost|Failed) before a call to ask(). In that case a answer was
registered as expected and never cancelled, which lead to a frozen client.

- Split _connectToPrimaryNode
- Move ConnectionClosed at generic level

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2357 71dcc9de-d417-0410-9af5-da40c76e7ee4

This is required to export data from NEO as done by FileStorage.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2356 71dcc9de-d417-0410-9af5-da40c76e7ee4

Same work was done in common and master specific partition table.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2355 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2354 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2353 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2352 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@2351 71dcc9de-d417-0410-9af5-da40c76e7ee4