Updates:
Status: Fixed
Comment #3 on issue 1115 by [email protected]: When doing master failover,
there is a race between starting WConfd and updating the configuration
https://code.google.com/p/ganeti/issues/detail?id=1115
Fixed in (for WConfd as well as Metad):
commit 7e16570b6de06c4ade8735c03b0d4af786c4bd46
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 15:27:13 2015 +0200
When connecting to Metad fails, log the full stack trace
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit f2d96bc3defdd541f8ad9fcd5508c886b5c7c7f1
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 15:11:00 2015 +0200
Set up the Metad client with allow_non_master
.. since the communication takes place on non-master nodes.
This ensures the client properly retries if there is a communication
failure.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit aa27b56a086d2da33f3e9eb1c9893c4f2d0a7da9
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 11:29:28 2015 +0200
Set up the configuration client properly on non-masters
If the configuration client is opened in the 'accept_foreign' mode,
meaning it is running on a non-master node temporarily, the option
needs to be propagated to the RPC client as well.
This fixes issue #1115.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit f828b17a7166e1261dd44da1d255811d94806201
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 11:28:41 2015 +0200
Add the 'allow_non_master' option to the WConfd RPC client
While at it, fix the call to the AbstractStubClient to properly pass the
keyword arguments.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit 706952c1759b2776d7c64ddaaa496b2860cfb5c4
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 11:26:18 2015 +0200
Add the option to disable master checks to the RPC client
The option is propagated to the Transport class and allows to disable
checks for the master node, if the client is run on a different node on
purpose.
While at it, fix the documentation for the arguments of the constructors
of the classes.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit 11efe0ff5c8b28dffe02fbeb84b8587e90a058d8
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 13:51:34 2015 +0200
Add 'allow_non_master' to the Luxi test transport class too
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit 17c0f0d47a12420c13f20b85825f7b73f7fa77c2
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 11:27:13 2015 +0200
Add 'allow_non_master' to FdTransport for compatibility
Since it serves as an alternative to the Transport class, it should
support the same constructor options, even if it doesn't use them.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit 1d464c1ce1b6d746473a4bbd42c4e94af08a0d62
Author: Petr Pudlak <[email protected]>
Date: Thu Jul 2 11:58:02 2015 +0200
Properly document all constructor arguments of Transport
.. while documenting allow_non_master.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
commit fa255d206e2c8a0acaf52f12c238b396ee1fe039
Author: Petr Pudlak <[email protected]>
Date: Fri Jun 5 14:13:48 2015 +0200
Allow the Transport class to be used for non-master nodes
If a communication failure occurred and the caller was not running on
the master node, Transport assumed that this itself was the cause of
the error condition.
However, for communication with the metadata daemon we need to support
non-master nodes as well.
Add a parameter that allows to use the class on non-master nodes.
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
Cherry-picked-from: ade70feb258a57ae0565395ba48ac2b3ef02b1c0
Signed-off-by: Petr Pudlak <[email protected]>
Reviewed-by: Helga Velroyen <[email protected]>
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
