-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Parity 1.7.6 - 1.8.0 stuck syncing #6787
Comments
probably a network issue. what's your configuration? |
I can reproduce this with version 1.7.7 Interesting though is: |
Issue is also present is 1.9.0-unstable Configuration is:
(--force-sealing and --allow-ips public was added in order to try to fix the issue, but does not seem to have helped) |
Here's a sample from the -lsync=trace output once parity has basically lost all peers and is no longer syncing blocks:
|
We got a fix from one of our colleagues, we are running with the following configs (in a docker container):
In {{chain}} we use kovan for testing |
What part of that specifically was the fix?... i don't see how adding |
Hi, my bad, should have posted the original:
We removed the changed ports to use I hope that helps |
I'm not using --no-wrap or docker, so still looking for a fix. |
Also I've tried running the connection through a VPN (in case there was some issue with the ISP), but still same issue. |
@diegopeleteiro that's an unrelated issue. @AdvancedStyle could you remove |
I've deleted out the nodes.json and tried again but the number of peers is still really low. Seems to start out at about 8/25 and then eventually over time drops to 1 or 2/25 peers at which point it starts to run into problems keeping synced (i'm guess just because the 1 remaining peer is not providing the data is a timely manner). Restarting parity syncs it back immediately due to initially be able to connect to 7 or 8 peers. Is there a way I force more peer connections? Note: ports 30301-30303 are forwarded at the router, but adding/removing this port forwarding doesn't really seem to make much of a difference as I guess it's utilizing the UPnP I wasn't running into these issues until after the 4370000 hardfork, and had been running 1.7.6 for a few days before for the hardfork without an issue. |
I notice that the "Public node URL" is the local network IP (enode://3f4cf050...9e2e@192.168.1.149:30303 Is this an issue? Should it be showing the public IP? |
Just throwing ideas out there.... Is it possible the node is doing something to get banned/blocked by peers? It is sending out quite a lot of transactions (probably average 3 or 4 per minute). And have a script that re-submits the raw transaction a couple of times to make sure it sticks (have had issues with transactions not propagating and just disappearing). Also node has a lot of keys (ie about 150,000+) however I don't see that this should have any effect on syncing with peers...but worth mentioning. Disk IO and memory usage on the server seem low/normal. |
The issue seems to somehow be related to reorgs and received conflicting blocks. For example look at the console output below. Was running along with 11 peers, then I received 2 versions of block #4386288 with different hashes (first one being the valid one). After that syncing seemed to freeze for a while and my peer count drop in half:
|
Looks like connections just timeout because all the IO handlers are busy enumerating account storage. Currently miner just reads all accounts from disk on every incoming transaction. This should be cached. |
Hello, i'm stuck too (and after many restart)
and i'm still stuck on block 2390979 :-/
I've tried to remove node.json, but same issue |
@jeremie-H that's a different issue. are you on hdd? |
@5chdn Hello ! yes, i'm on hdd |
@jeremie-H You are probably experiencing #6280 - Try to get an SSD or try a warp-sync. |
I have same issue and I think this issue is not related with insufficient I/O or CPU resources.
my environment is :
tried clear db and @5chdn can you reopen this issue? |
@5chdn I have similar problems, which 1.7.x version should I downgrade to ? (I'm using 1.7.12 right now) |
@pgrzesik 1.7.11 but note that 1.7.x will reach the end of life very soon. |
Thanks @5chdn ! |
@5chdn thank you! |
After the fork I'm having issues with parity getting stuck in "syncing" status. I've tested in both 1.7.6 and 1.8.0 and the issue happens randomly after running fine for about 30 minutes. All of a sudden the peers dropped from 4 to 1 and then stuck in "syncing" the same block.
Here's an example of the parity output where the "Syncing" occurs.
The text was updated successfully, but these errors were encountered: