Since upgrading my proxy node on my Raspberry Pi to v0.4.4.6-nvc-12-gbafc8da-beta, I'm observing a behaviour I haven't seen before in debug.log:
getblocks 67859 to 00000000000000000000 limit 500
getblocks stopping at limit 68358 d537d885299074f8efb1
getblocks 68359 to 00000000000000000000 limit 500
getblocks stopping at limit 68858 000000000007acf2049f
getblocks 68859 to 00000000000000000000 limit 500
getblocks stopping at limit 69358 b0d680ecdd0e4d6984cc
getblocks 65859 to 1d8fafa34a7f7d29684f limit 500
getblocks stopping at limit 66358 e7154a17ebfe38e3797a
getblocks 66359 to 00000000000000000000 limit 500
getblocks stopping at limit 66858 000000000030a21c6998
getblocks 66859 to 00000000000000000000 limit 500
getblocks stopping at limit 67358 e247514923cb9454c4af
getblocks 67359 to 00000000000000000000 limit 500
getblocks stopping at limit 67858 00000000006d77fcb0ab
getblocks 67859 to 00000000000000000000 limit 500
getblocks stopping at limit 68358 d537d885299074f8efb1
getblocks 68359 to 00000000000000000000 limit 500
getblocks stopping at limit 68858 000000000007acf2049f
getblocks 68859 to 00000000000000000000 limit 500
getblocks stopping at limit 69358 b0d680ecdd0e4d6984cc
You can see the getblocks looping, i.e. going up to 68859, then starting again at 65859. Since unfortunately it doesn't log IP I cannot see if it's always the same node requesting this.
When stopping & restarting the node , it stays normal for several days until I start seeing this again(happened about 5 times so far). Otherwise, the node seems to remain functional, i.e. forwarding blocks etc.
Anyone else seen this behaviour?
I seen that many times on my cubie and I see this now on my VDS.
I don't think that it's the same node, I think this requests are coming from buggy nodes (< 0.4.4.5). The buggy client is able to see that network is on the longer chain currently, but unable to accept this blocks. That's why it tries to sync again and again with a randomly chosen node, bans it and then tries to sync from another one. This can cause some traffic or CPU overhead on normal node, but overall, it's harmless.
We'll add a special condition for incoming connections initiated from versions prior to 0.4.4.5.
You can use a small patch to resolve this by yourself:
https://github.com/novacoin-project/novacoin/pull/2bizzargin