how long does it take for a node to join and catch up?

[dckc]

started:

2020-02-29 05:47:26,392 [main] INFO  coop.rchain.node.Main$ - RChain Node 0.9.22 (b630ecb37b4d0c4e71df7c06aa0dbfa98f761255)

2020-02-29 09:27:43,680 [tl-dispatcher-server-queue-130] INFO  coop.rchain.casper.engine.Engine$ - Making a transition to Running state.

[dckc]

looks like it won't ever catch up.

Mainnet seems to be making a new block every 30 sec, and my node is spending about that much time between "Computed parents post state for Block #1425 (63292b82d0...)..." events. Mainnet has over 8400 blocks.

rnode kc-strip perf notes - Google Sheets

Processor	2x Opteron 6128 HE
Ram	32 GiB
Bandwidth	1Gbit Port: 100TB Transfer
Disk	2 TB

[dckc]

my node is now at 3467; mainnet is at 10221

@zsluedem offered to look at my logs, so I shared them with him:

~/projects/rstake/logs$ ls -lsh
total 43M
 15M -rw-r--r-- 1 connolly connolly  15M Mar  1 09:20 rnode.log.gz
3.9M -rw-rw-r-- 1 connolly connolly 3.9M Mar  1 09:25 rnode.log.gz.part-aa
3.9M -rw-rw-r-- 1 connolly connolly 3.9M Mar  1 09:25 rnode.log.gz.part-ab
3.9M -rw-rw-r-- 1 connolly connolly 3.9M Mar  1 09:25 rnode.log.gz.part-ac
2.7M -rw-rw-r-- 1 connolly connolly 2.7M Mar  1 09:25 rnode.log.gz.part-ad

[zsluedem]

Just want to let you know that we found the bug and have a solution for that. Hopefully we can release a new version next week and resole the problem.

[dckc]

looks like https://github.com/rchain/rchain/pull/2902 is the fix. I'm almost tempted to build it myself... but I'm mostly using docker and I'm not familiar with that part of the build process.

mainnet is at block 25055 ; my node is at 20092. it's pretty clearly not going to catch up without a software fix. I think I'll shut it down for now.

[nzpr]

@dckc if you still want to build, this tutorial should be up to date and working https://github.com/rchain/rchain/blob/dev/DEVELOPER.md But instead of building bnfc which is not fast you can try just put this binary on the PATH https://github.com/nzpr/rnode-parallelism-test/blob/master/bnfc

[dckc]

I see @jimscarver and co have it deployed: "We have run nuttzipper/rnode:dev for three days and are up to about 14600 of 20550 testnet block"

$ docker pull nuttzipper/rnode:dev
...
Digest: sha256:e9e0a1bf896be4ebd8b0a6b535de3e9b2f172b270df6b1d777b73c81746399b7

[dckc]

fixed in 0.9.23? https://github.com/rchain/rchain/releases/tag/v0.9.23 d6e2c31, reportedly

@fabcotech says his node is catching up on the testnet, though slowly... he gets "seqNum":0 at /api/blocks/1 vs "seqNum":10751 at https://observer.testnet.rchain.coop/api/blocks/1

I'll stand by until I see either something clear in the release notes or experimental verification.