JoelKatz
04d48a6d21
Emergency fix: Sanely handle taking a position we previously failed to acquire.
2013-04-05 09:57:24 -07:00
JoelKatz
7131564eb1
Emergency patch to fix ledger not closing.
...
Do not timeout on acquiring a transaction set if a trusted peer
still proposes it. Instead, fetch more aggressively.
2013-04-05 09:41:43 -07:00
JoelKatz
cd14be1afb
Better logging of untrusted validations. (To make sure there's no bug.)
2013-03-21 06:03:47 -07:00
JoelKatz
371d52cc9e
Move LCT code to the right file.
...
Be smarter in the consensus set filter code.
2013-03-20 09:22:30 -07:00
JoelKatz
6e9a1f09b5
Break TransacionAcquire into its own implementation file.
2013-03-20 08:54:17 -07:00
JoelKatz
d4323f2ef2
Fix a case where a node that closed slightly ahead of other ledgers can
...
think it's out of sync with the network and cause it to jump backwards one
ledger, causing it to be out of sync because the network was about to catch
up to it.
2013-03-14 11:48:29 -07:00
JoelKatz
e56f7a8689
Fix a rare race condition.
2013-03-14 09:48:43 -07:00
JoelKatz
774c38f99c
Fix tie breaking in ledger validation during preClose consensus phase.
2013-03-11 15:30:55 -07:00
JoelKatz
d1169ee86d
Make sure all load sources are named.
2013-03-04 16:22:09 -08:00
JoelKatz
a3dcc36e05
Typo and some extra debug.
2013-03-03 14:07:46 -08:00
JoelKatz
4b97f422f3
Rework how we declare a close time consensus to prevent stalls.
2013-03-01 13:46:13 -08:00
JoelKatz
92c1766f49
Ack! A delayed consensus causes the close time threshold to ratchet the wrong way.
2013-03-01 13:20:08 -08:00
JoelKatz
5089db1518
Add some extra debug logging to track absence of CT consensus.
2013-03-01 13:17:14 -08:00
JoelKatz
85cd5bc47f
Fix the bug that crashed ec2. Must hold a TransactionAcquire::pointer
...
to call a function that calls erase.
2013-02-28 19:41:03 -08:00
JoelKatz
7b14d2d44a
Fig a bug in the way we traverse TaggedCaches and a similar bug in mPeerPositions.
2013-02-25 20:28:20 -08:00
JoelKatz
076add0e73
Be more aggressive about relaying disputed transactions.
2013-02-24 14:13:20 -08:00
JoelKatz
92be1c70f5
Hopefully prevent a jump forward at the end of the consensus process.
...
If not, at least figure out exactly where in the consensus state machine that's happening.
2013-02-24 13:47:21 -08:00
JoelKatz
b512ed3df2
Correctly handle local failures in the consensus transaction application process.
...
Local failures are not retriable.
2013-02-11 16:43:29 -08:00
JoelKatz
8653d679b3
Change rule slightly if consensus seems stuck.
2013-02-10 13:23:11 -08:00
JoelKatz
b39a977282
Cleanups.
2013-02-10 01:09:37 -08:00
JoelKatz
f8f56b96e0
Tiny cleanups.
2013-02-09 20:07:53 -08:00
JoelKatz
4ef325e8a9
More logging improvements.
2013-02-09 16:16:46 -08:00
JoelKatz
7ea9ad9de7
Improve consensus logging.
2013-02-09 15:41:21 -08:00
JoelKatz
79cbb0eea0
Cleanups.
2013-02-09 15:34:43 -08:00
JoelKatz
18167be5e3
Add new 'consensus_info' RPC command to help troubleshoot consensus convergence issues.
2013-02-09 14:21:57 -08:00
JoelKatz
517c1c48ff
Clean up pass counts.
2013-01-21 18:26:24 -08:00
JoelKatz
18001a0058
Fix retry bug.
2013-01-21 17:55:39 -08:00
JoelKatz
6e746c5851
Fix a harmless bug that caused transactions that succeded early to be retried once.
2013-01-20 21:59:34 -08:00
JoelKatz
07091cfc3c
Fixes.
2013-01-19 14:31:31 -08:00
JoelKatz
4c6920dd55
Add extra debug to better understand how the txn retry logic is working.
...
Avoid an extra transaction pass caused by failed transactions counting as changes.
Downgrade some debug messages from INFO to DEBUG.
2013-01-19 14:09:42 -08:00
JoelKatz
80d98e55a9
Clean up JSON reporting of booleans and network state.
2013-01-18 17:38:11 -08:00
JoelKatz
79d1727b38
Hopefully, handle partial success correctly. Retry engine.
2013-01-18 10:17:21 -08:00
JoelKatz
1acdad8601
Make it compile.
2013-01-09 09:38:41 -08:00
JoelKatz
eadaaaa0fd
Clean up debug.
2013-01-09 09:22:25 -08:00
JoelKatz
bbd32faf05
Demote some older logs.
2013-01-08 14:50:45 -08:00
JoelKatz
7a2f098ac4
Reduce some chatty logging.
2013-01-07 23:47:21 -08:00
JoelKatz
c2a4cc8321
Get rid of transaction acquire entries ASAP.
2013-01-07 14:44:18 -08:00
JoelKatz
63fd3818fb
Rather than counting PeerSet's, count LedgerAcquire's and TransactionAcquire's.
2013-01-06 17:32:25 -08:00
JoelKatz
ee89904a00
Cleanup trigger.
2013-01-04 20:26:26 -08:00
JoelKatz
11cb109c16
Another safety.
2013-01-03 23:23:10 -08:00
JoelKatz
6849c579eb
Extra safety.
2013-01-03 23:12:48 -08:00
JoelKatz
0d8ed85bd4
Update.
2013-01-03 22:44:56 -08:00
JoelKatz
e78b5d11b6
Bugfix.
2013-01-03 22:33:55 -08:00
JoelKatz
d4f4d9bf78
Redesign the way the acquire timer is (re)set so that we won't have bugs
...
where we fail to arm it.
2013-01-03 21:25:40 -08:00
JoelKatz
40b87e9157
Fix a bug that would cause the server to exit the "need network ledger" state prematurely.
2012-12-30 23:48:53 -08:00
JoelKatz
a169167030
Fix a bug where a node doesn't get a chance to finish acquiring a TX set
...
before all nodes forget it because they're done with it, leaving a node
behind the consensus.
2012-12-27 20:25:12 -08:00
JoelKatz
7c04eded0f
Improve some consensus logging.
2012-12-26 22:30:18 -08:00
JoelKatz
79d139e2ce
Don't let a tranasction set acquire stall.
2012-12-25 23:00:39 -08:00
JoelKatz
9689f94f5a
Cleanups.
2012-12-23 17:52:34 -08:00
JoelKatz
2c535940ac
Make the transaction engine report whether it added the transaction.
2012-12-23 17:42:04 -08:00
