- Improvements on the performance of communication system
- big clean up of the code (many old methods deleted)
- corrected bug in out of context (now, a consensus is only started if a propose is received)
Note: still need to verify if state transfer and leader change works right

In this commit, it was fixed a bug on the comuniation system that would make the clients freeze  after the removal of a replica

In this commit, I fixed a bug that manifested itself during the leader change protocol (when the original leader crashed). I also added a few more locks to protect the sessionTable of the comunication system.

In this commit, I have submitted the code for the leader change. Unfortunately there is some kind of conflict between the code I developed for an older version of smart, and the most recent code available on the svn. Such conflict renders the leader change protocol faulty. Further debugging is necessary. The normal operating remains good, though.

Small modifications and corrections on the counter demo plus the creation of MultiCounterClient that starts N JVMs with counterclients redirecting their outputs different files. I did that to try to catch the error Andre reported, but it doesn't appear on this example.

Corrected small bugs related with number of messages needed for non-read-only request (it is 2f+1 instaead of f+1) and also added a invokeTimeout property that defines the amount of time a client wait for replies from an invoke before return null.

Improved code for shuting down the client if the servers are not available... additionally, exchanged Hastables by HashMaps

store replies to be sent if requests arrive from client after their processing (when comming from the leader). Very subtle missing feature found by Andre!!!

No commit message

No commit message

No commit message

No commit message

No commit message

No commit message

No commit message

No commit message

Added the patch produced by Andre Nogueira, that fixes the case of when a replica joins the group, the information about its ip address and port not being broadcasted to the other replicas.

No commit message

Fixed a bug on the communication system that could make the virtual machine consume way too much CPU.

Fixed a bug in ExecutionManager.java where not all stoppedMsgs would be removed from the List upon restart

In this commit, it has been added a timeout for the state transfer protocol, and fixed an implementantion error regarding f + 1 state hashes consistent amoung them, and a complete state inconsistent with those ashes.

It was also fixed a bug that would crash clients when a a reconfiguration was applied, and also fixed 2 bugs related to the checkpoints, which  would make replicas crash.

Finally, there are also some lines of code that have been slightly modified, but which shouls not affect the behavior of SMaRt.

Not much was changed. This commit only has minor changes, a bug fix that covers a recovered old leader so it doesn't trigger unnecessary timeouts, and some code to help test the library

In this commit, it was added a new demo, a bug fix for the TTP, and also a new method for in the ServiceReplica class that makes the replica leave the group.

Removed out of context thread. The processing out of context messages is now made at the delivery thread and Server CS thread
Removed paxos internal timers.  From now on, the external timers from the total ordering layer shall trigger the leader change
Added a few debugging messages for delivering messages
Fixed a bug related to signatures of forwarded messages from replicas to the leader
Fixed a bug on the TTP that would make it crash in some cases
Fixed 2 bugs on the state transfer protocol responsable for not letting the old leader fetch the state once it recovered

Include session on the messages... corrected the problem of clients block when they ressurrect (so, now a client can run multiple times with the same id)

In this commit, I added the rest of the files that are required to run SMaRt, but are not part of the code.

In this commit, I added the rest of the files that are required to run SMaRt, but are not part of the code.

In this commit, it was added a new feature to SMaRt. It now supports dynamic views. Some bugs from the previous commit were also fixed. This was a fairly extensive update to SMaRt.

Nonces are now generated for each TOMMessage in a Batch with a single seed that is sent along with the batch instead of a fixed number of nonces per Batch.
Decided notification in Consensus.java now is done via a sync monitor instead of busy waiting on the decided value to be set. Because of the Java MemoryModel the former mechanism was problematic and could theoretically lead to the DeliveryThread never learning the new value - which is very unlikely though.