v0.10.16 - Onyx v2

flare

Administrator
Dash Core Team
Moderator
May 18, 2014
2,286
2,404
1,183
Germany
Hmmmm.....
How about this?
Code:
cat ~/.darkcoin/debug.log | grep "CDarkSendPool::GetMasterNodeVin" | tail
Checked this against ALL my customers nodes, i don't get a single entry in any debug.log ...
 

jiggytom

Member
Nov 14, 2014
104
22
68
wonder what's up with these..says inactive for active duration, but status active.
Masternode IP:portPort Check (Next check)VersionMasternode PubkeyBalanceActive DurationLast SeenCountryStatus
(Active Score)
104.131.134.41:9999Open (42m52s)0.10.16.11XtfLy5bx2WUN43Uto2Hq2WozAVhF5ShtWE[1][2]1,000.000Inactive6m48s
United StatesActive (100%)
104.131.236.114:9999Open (42m16s)0.10.16.11XdAdrkjBbSMSe8ofVtS7S4PPQXzFbBwsSi[1][2]1,000.000Inactive6m47s
United StatesActive (100%)
107.170.13.222:9999Open (42m17s)0.10.16.11XsmWeGdR6ZQLZUJRi7ut16keKMARssNZVy[1][2]1,000.000Inactive6m46s
United StatesActive (100%)
107.170.157.166:9999Open (42m17s)0.10.16.11XtcEBhogcKvheZ41hAL3HLVeEAcWpED13z[1][2]1,000.000Inactive6m45s
United StatesActive (100%)
107.170.196.35:9999Open (42m17s)0.10.16.11XkQnk5U9fUWhvrnn4uKPGZn3wj3UKxvuhX[1][2]1,000.000Inactive6m48s
United StatesActive (100%)
 

GermanRed+

Active Member
Aug 28, 2014
299
109
113
Masternodes still getting inactive after 60 minutes, even with v0.10.16.11.

A script automatically starts them again once they are down, but it's still annoying.

This is how my log looks like for the last 30 hours or so, the big dip in the middle was the update to v0.10.16.11 last night:
View attachment 630
I also have a script checking activity of MN's and winners. After I stopped running the script, the problem goes away. Do you want to try stop running your script? There may be some bug that kicks the MN off the list after certain query limit.

EDIT: Have been running without being kicked off the network for six hours now. The script worked perfectly with 0.10.15.21. So, I'm sure it is not the script that is buggy. I also wait 10 seconds after stop and 60+ seconds after start.
 
Last edited by a moderator:

flare

Administrator
Dash Core Team
Moderator
May 18, 2014
2,286
2,404
1,183
Germany
UdjinM6

I checked for "CActiveMasternode::RegisterAsMasterNode() - Masternode input = " on all nodes now, and i am getting a certain percentage of hits: ~ 20% of all nodes have these entries.
Seems to be a race condition in the masternode startup code...
 

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183
Last edited by a moderator:

flare

Administrator
Dash Core Team
Moderator
May 18, 2014
2,286
2,404
1,183
Germany
Only one way for them to stay online in this situation - they are "in the future".

We need a check like this https://github.com/darkcoin/darkcoin/blob/master/src/masternode.cpp#L178-L181 to be implemented in CMasterNode::Check() https://github.com/darkcoin/darkcoin/blob/master/src/masternode.cpp#L464 and maybe somewhere in dsee message processing.

EDIT: fixed -->Check
I've checked the time, both masternodes and local laptop - all in sync, despite being different timezones.
 

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183
I've checked the time, both masternodes and local laptop - all in sync, despite being different timezones.
That's confusing me cause I have no more ideas..:oops:
The last one. From somewhere else (not from the node you check):
Code:
darkcoind masternode list lastseen | grep "IP_OF_SOME_NODE"
What is the unix time you get for this? Is it something in the last hour or so?
http://www.onlineconversion.com/unix_time.htm
 

Light

Well-known Member
Foundation Member
Jun 4, 2014
346
256
233
Sync problems with v101611 anybody?
Seems ok for now.
I had before on Mac but after deleting peer.dat it was solved
 
Last edited by a moderator:

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183
Sync problems with v101611 anybody?
Seems ok for now.
My qt wallet stuck today at "1 hour behind". Closed, opened it again and it synced.
I guess Evan will start to squash some old bugs once he is satisfied how new Darksend implementation works :)

EDIT: I didn't remove peers.dat.... will do this now
 

AjM

Well-known Member
Foundation Member
Jun 23, 2014
1,341
575
283
Finland
I guess Evan will start to squash some old bugs once he is satisfied how new Darksend implementation works :)

EDIT: I didn't remove peers.dat.... will do this now
Cool, is this rumor or fact?
 

crowning

Well-known Member
May 29, 2014
1,414
1,997
183
Alpha Centauri Bc
Checked this against ALL my customers nodes, i don't get a single entry in any debug.log ...
Me neither.

But what's REALLY odd:
Code:
[email protected]:~$ cat .darkcoin/debug.log | grep "CActiveMasternode" | tail -n 20
I don't get ONE single CActiveMasternode::RegisterAsMasterNode() entry in the logs of 3 of my Masternodes.
Nothing. Not a single one.
And there should plenty of them, because these Masternodes are listed as active.

I'm using the official Linux 64 bit binaries.

I think I'll build them myself from GIT like you did, maybe there's a difference...
 
  • Like
Reactions: flare and UdjinM6

djcrypto

Member
May 27, 2014
180
94
88
My qt wallet stuck today at "1 hour behind". Closed, opened it again and it synced.
I guess Evan will start to squash some old bugs once he is satisfied how new Darksend implementation works :)

EDIT: I didn't remove peers.dat.... will do this now
Are there other instructions to follow when updating to 16.11?
Do I remove peers.dat on both local and server?
 

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183
Are there other instructions to follow when updating to 16.11?
Do I remove peers.dat on both local and server?
There is no really need for that I believe, it should just help to avoid "old" clients so wallet will not try to connect to them, fail :) and remove them from the list one by one.
But in a matter of testing and looking in debug.log to find something unusal this might help. But I see nothing unusual so far....
 
  • Like
Reactions: djcrypto

AjM

Well-known Member
Foundation Member
Jun 23, 2014
1,341
575
283
Finland
Are there other instructions to follow when updating to 16.11?
Do I remove peers.dat on both local and server?
You can do that, no harm, just in case, i always delete peers.dat from both local and remote when updating.
 

TanteStefana

Grizzled Member
Foundation Member
Mar 9, 2014
2,871
1,863
1,283
Oh, so this is where ya'all are talking ;)

Has Evan uploaded 10.16.12 yet? He said he fixed mixing issues on BTC this morning...
 

crowning

Well-known Member
May 29, 2014
1,414
1,997
183
Alpha Centauri Bc
Checked this against ALL my customers nodes, i don't get a single entry in any debug.log ...
I think I'll add some lines of additional Log, build again and see what's going on...
Okay, adding a ton of my own debug info tracked this one down, and even when I'm a bit ashamed to admit it's user error the implications of my findings are too interesting to not share them here:

  1. I have a hot/cold setup for my Masternodes.
  2. I have a script which updates my remote Masternodes automatically. The binary, a generated darkcoin.conf and a couple of other things. Somehow :) about one month ago the configuration entry "masternode=1 got lost, so from that day on all of my hot Masternodes did NOT have that line in darkcoin.conf. :facepalm: (the cold ones have, though)
  3. This means CActiveMasternode::RegisterAsMasterNode() is never successdully executed:
    1. Private key is not checked
    2. External IP and port is not checked
    3. Wallet lock is not checked
    4. vin is not checked
    5. Masternode is not relayed to the peers
    6. A ton of other things...
  4. Most of the above is not really THAT important for a hot Masternode to run properly , but I still fail to see why my Masternodes were seen (and paid of course) by the rest of the network at all.
 
  • Like
Reactions: UdjinM6

oblox

Well-known Member
Aug 6, 2014
1,032
537
183
Anyone not using superputty or some other multi-instance update to update multiple nodes is wasting precious time. Makes it incredibly easy to update.
 

darkwing

Active Member
Apr 8, 2014
149
110
103
denom working :) well done evan.

*Edit* have been hit with two 0.1 fees

*edit2* 3 fees and only 1 round complete

Will turn off and wait a few days

*Edit 3* I see success reported on bct and others are successful.
So guess I've just been really unlucky. 3 fees/2rounds
 
Last edited by a moderator:

UdjinM6

Official Dash Dev
Core Developer
Dash Core Team
May 20, 2014
3,639
3,537
1,183
Okay, adding a ton of my own debug info tracked this one down, and even when I'm a bit ashamed to admit it's user error the implications of my findings are too interesting to not share them here:

  1. I have a hot/cold setup for my Masternodes.
  2. I have a script which updates my remote Masternodes automatically. The binary, a generated darkcoin.conf and a couple of other things. Somehow :) about one month ago the configuration entry "masternode=1 got lost, so from that day on all of my hot Masternodes did NOT have that line in darkcoin.conf. :facepalm: (the cold ones have, though)
  3. This means CActiveMasternode::RegisterAsMasterNode() is never successdully executed:
    1. Private key is not checked
    2. External IP and port is not checked
    3. Wallet lock is not checked
    4. vin is not checked
    5. Masternode is not relayed to the peers
    6. A ton of other things...
  4. Most of the above is not really THAT important for a hot Masternode to run properly , but I still fail to see why my Masternodes were seen (and paid of course) by the rest of the network at all.
There are some checks of vin in ProcessMessageMasternode (on receiver side) and in ThreadCheckDarkSendPool actually so it should be ok.

When there is no masternode=1 on remote node it will check keys and vin https://github.com/darkcoin/darkcoin/blob/master/src/masternode.cpp#L129
update lastTimeSeen and add itself into MN list https://github.com/darkcoin/darkcoin/blob/master/src/masternode.cpp#L141-L142
and only then it will fail https://github.com/darkcoin/darkcoin/blob/master/src/activemasternode.cpp#L296
but still relay itself to others https://github.com/darkcoin/darkcoin/blob/master/src/masternode.cpp#L149
So you'll get your 70 minutes and then your node will become inactive.
Definitely should fail faster :)

EDIT: http://jira.darkcoin.qa/browse/DRK-125
 
Last edited by a moderator: