Continuing the discussion from 0.17.1rc1 reindex-chainstate, assumevalid=0:
So this was running 0.17.1rc1 with a full reindex, assumevalid=0, and a full txindex in place. Bitcoingold.conf was:
[main] debug=1 assumevalid=0 reindex=1 txindex=1
The start of the run:
2019-09-29T15:30:16Z 2019-09-29T15:30:16Z Bitcoin Gold version v0.17.1rc1 (release build) 2019-09-29T15:30:16Z InitParameterInteraction: parameter interaction: -whitelistforcerelay=1 -> setting -whitelistrelay=1 2019-09-29T15:30:18Z Validating signatures for all blocks.
Chain sync was finished and up-to-date here:
2019-09-30T04:08:16Z UpdateTip: new best=000000019f550bf1039de36e359d2f13697fe1f5d35430222f4382f2650447a6 height=601193 version=0x20000000 log2_work=87.340668 tx=266527475 date='2019-09-30T03:35:43Z' progress=0.634711 cache=228.9MiB(1402671txo)
For a run time of about 12:40 (12 hours 40 minutes.
This is good news! I can confirm that the rebuild error when running a
assumevalid=0 is gone under Windows. This is relevant because I recently (falsely) reported that rebuilds were failing, but that turned out to be because of a couple of bad sectors on my own hard drive.
In the interest of thoroughness, I ran several rebuilds with
asssumevalid=0, both with
reindex-chainstate without and with a fully build txindex; in all situations, the rebuild works and the bug is gone.