This repository has been archived by the owner on Oct 4, 2019. It is now read-only.
Stops synchronization after some time #398
Assignees
Labels
Comments
|
I am unable to reproduce. Have synced Morden fully 3 times today. How long are you waiting when sync stops? What does it look like? Do you have logs from when this happens? Can you share your docker/k8s config so I can try with that? |
|
I synced Morden recently few times on different machines (both virtual and physical), with no problems. In the past I have some pauses in synchronization - in the morning I found that nothing was logged for about 3 hours in the night. But I ignored the issue, because geth resumed automatically, I was not sure about the stability of network link. |
|
It may also be possible that this is related to an ECIP1017 fork. Many nodes are not yet upgraded to include the change on Morden, eg our PR to Parity merged only 3 weeks ago - and many geth nodes I see are still running |
|
@whilei that was happening even for initial blocks, say before 1500000, so a potential incompatibility after block 2M should not affect that |
|
Ok, let's keep it in mind, we'll get numbers as soon as we'll get better monitoring/logging/console |
|
Possibly rel #346 |
whilei
added a commit
that referenced
this issue
Nov 23, 2017
… re: reorg solution: modify reorg conditional in blockchain insertchain to use smaller block number or randomized block reorg based on http://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf rel #346 rel #398 possibly #378
whilei
added a commit
to whilei/go-ethereum
that referenced
this issue
Nov 24, 2017
… re: reorg solution: modify reorg conditional in blockchain insertchain to use smaller block number or randomized block reorg based on http://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf rel ethereumproject#346 rel ethereumproject#398 possibly ethereumproject#378
whilei
added a commit
that referenced
this issue
Nov 24, 2017
… re: reorg solution: modify reorg conditional in blockchain insertchain to use smaller block number or randomized block reorg based on http://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf rel #346 rel #398 possibly #378
sorpaas
pushed a commit
that referenced
this issue
Dec 7, 2017
… re: reorg solution: modify reorg conditional in blockchain insertchain to use smaller block number or randomized block reorg based on http://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf rel #346 rel #398 possibly #378
|
Still happening with 4.2: Notice this |
|
Possibly related to #448, since this is an |
|
Another example of this issue has been brought to my attention: Detailed logs from apparent time of failure: |
|
I have again reproduced this. Looking like Also, just to note this bug is not limited to Morden. Above two logs from mainnet, for example. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Geth 4.1.0, Ubuntu+Docker+Kubernetes, Morden chain
Always stops chain synchronization after 1-4 hours, even for initial sync. Successfully continues download after restart.
Maybe related lack of peers on morden. It seems that geth removes a peer from the list, because of some error, and at the end gets left without any peer at all, even if most of them are back online.
The text was updated successfully, but these errors were encountered: