[spam][personal][crazy] Trying to Build Bitcoin from Source

[Karl]

0637 edt
this error is from bitcoinsv: bitcoind:
/home/user/src/bitcoin-sv/src/validation.cpp:2889: bool
BlockConnector::Connect(cons
t task::CCancellationToken&): Assertion `hashPrevBlock ==
view.GetBestBlock()' failed.
Aborted (core dumped)

i think it is innocuous and just looks scary, unsure.  reindexing and
wiping chainstate pushed it away.

grarpamp recommended an incredible idea of indexing the files on a
blockchain.  doing this would work best if i had a synced chain.
bitcoin is presently on 2015-05 so is still just barely started
syncing.  maybe i can find a downloadable chain somewhere.  0638

0639 and i found
https://bitcoin.stackexchange.com/questions/799/can-i-download-the-whole-block-chain-from-somewhere

it contains this sentence: " Network and synchronization code have
been immensely improved over the years so that the initial block
download via the network is usually acceptably fast."  This is not my
experience at all.  maybe the db flags grarpamp mentioned are my
issue?  i don't know.

i likely also have some blocks on harddrives indoors.  indoors is
dangerous for me because it is hard to leave.  0640

it's 0642 and my internet does not appear to be working.  i'll inspect
the antenna.  the phone i'm using as a flashlight is at 8% after being
plugged in when i got here.  0645 and net is back.

it's 0647 and now i'm at https://bitcointalk.org/index.php?topic=145386.0

it's 0650 and i'm using google to search for "how to sync blockchain fast".

0655 and i'm trying to figure dbcache out.  it's in the bitcoind
--help.  dbbatchsize i don't see there yet.  i bumped mine from 450M
to 3*1024.  hasn't allocated that much ram yet.

0657 given this will take longer to complete than i can remember what
i am doing, options include:
- getting blockchains from indoors
- using another client, such as a light client
- doing something else

i've been thinking a little about libbtc.  it's at
https://github.com/libbtc/libbtc and i think i may unfortunately be a
maintainer of it.  it was a light client.  it doesn't have the bloom
protocol stuff implemented yet, but it can probably be used to scan
peers for blocks pretty reasonably, maybe even validate them using
some old protocol, who knows.

there's also libbitcoin.  the problem with libbitcoin is that it is
huge.  binaries take up a lot of disk space.  it also has a large
build infrastructure.  wouldn't normally be a problem if you weren't
me, i imagine.

i have an existing project for indexing chains.  i actually have a lot
of them.  but one of them is recent.  it's hard for me to access and
think about these projects, and i tend to make them a mess before
dropping them.  it's 0700.  let's see if i can find it.

additionally, it's best to have the whole chain, and that's hard for
me with the current amount of space i have.

my concern with buying more disks is that i tend to fill them up in
ways that are hard to manage.  they also develop corruption, and then
i preserve them in the hope i can recover my files at some point.  the
files are generally unreasonable to recover for me, but that's very
unfamiliar to me, which confuses things further.

learning zfs seems like less of an investment than working on a chain
project.  i've tried them both before, failed at them both before, but
chains keep growing whereas zfs isn't really that flexible.  it does
provide rollbacks which would really help with my corruption issues.
chains do too, but are less normal to use for storage.

filecoin btw is kind of a joke.  they rewrote the protocol so that
interactions with hosters are done out of band, and the protocol
doesn't enforce maximizing the storage economy the way a proof of work
does.  so basically you can't use the blockchain to escape file
censorship because the storage protocol is out of band and you're
responsible for doing your own duplication of content by finding
multiple hosts.

it didn't work for me.  hosters rejected my content or didn't respond
to my pings.

the previous whitepaper looked great.  but we have lots of other
chains and lots of ways to get people to respond to pings, if one
focuses on problems.  filecoin takes a modern system to run (uses
gpus), is another barrier.

it's 0705.  what is my last chain interface project.

confused.  another approach would be to contribute stuff to bitcoin so
it can be used as a light client.  that's probably sorely needed.  my
phone alarm is going off loudly.

it's 0706 and my body is trying to leave this system.  inner conflicts exist.

it's 0706 and two blockchains are syncing pruned. my system has 3G of
ram cached by the os, thinking my dbcache value could have been
larger, dunno.

0708 i'll look through my github repos to see if i can find anything
like a blockchain database interface

it's 0709 and my internet is down.  the device has disappeared from
network manager.

when i type nmtui-connect, the system hangs.  this is happens to me
sometimes.  this email is likely to be lost, so i'll copy-paste it
into a dated file, and keep writing in the file.

it's 0710 and i'm writing in a file instead of a web browser.  my
system is experiencing a kernel issue associated with my wifi device.
blockchains think they are syncing, but have no network to do so with.
some zstd processes are still running, compressing things.

it's 0711 and i'm switching into terminal mode

i'll use a different tty to try to become root.  this can take a long
time as the kernel waits for network time out, doing some host lookup
in sudo. [maybe, guessing.]

first i get to log in.  it's 0711.  i've entered my username in a tty
login, and am waiting for the password prompt to load.  this is how my
system behaves when my wifi device glitches out.
it's 0712 and the password prompt loaded.  i've logged in, and am
waiting for sudo to respond.  it's 0713.

this is a redhat system, a gifted subscription from a family member.
it's running redhat 7.  i've opened a support ticket around the wifi
issue before, but i ended up having cognitive issues before resolving
it.

one of the upshots of that various exploration, is that now crashdumps
are enabled in my kernel.  there's some way to issue some sysrq to
make one.  it makes sense to do this in this system state.

sudo is still waiting.  it's 0714 and i'm looking for information on
triggering crashdumps on my phone.

it's 0715 and sudo is still pending, stalled after typing the command
in.  i've loaded 'How to troubleshoot kernel crashes', a redhat
article, on my phone.
    after spending some time with redhat, i've also started pursuing
ubuntu, which seems to have a more widely used package management
system.  generally, if you have paid support, you have an issue:
        you need to be able to contact the people offering you paid
support.  this is most reliable if you can physically visit them.
haven't looked into that.

the article says you can write 'c' to /proc/sysrq-trigger.  maybe you
cam hit c as a key too, dunno.  it's 07:19.  sudo still pending.

it'd better stop my zstd processes so that i don't get confused by
having zst files with half content if the system reboots.
stopped, partial zst files deleted.  some of thse have been running all night.

i must not know how to do it.  when i hit alt-sysrq, some kernel stuff
scrolled by, then hitting c did nothing.

here's the kernel stuff:

[147513.443091] CPU: 2 PID: 0 Comm: swapper/2 Kdump: loaded Tainted: P
          OE  ------------   3.10.0-1160.45.1.el7.x86_64 #1
[147513.443219] Hardware name: LENOVO 20K70004US/20K70004US, BIOS
N1QET79W (1.54 ) 11/16/2018
[147513.443303] task: ffff89ec367a5280 ti: ffff89ec36204000 task.ti:
ffff89ec36204000
[147513.443379] RIP: 0010:[<ffffffff9fbc8dc7>]  [<ffffffff9fbc8dc7>]
cpuidle_enter_state+0x57/0xd0
[147513.443483] RSP: 0018:ffff89ec36207e60  EFLAGS: 00000206
[147513.443540] RAX: 00008629e94d4d1d RBX: 00000000000159a0 RCX:
0000000000000018
[147513.443648] RDX: 0000000225c17d03 RSI: ffff89ec36207fd8 RDI:
00008629e94d4d1d
[147513.443817] RBP: ffff89ec36207e88 R08: 0000000000000ae5 R09:
0000000000000018
[147513.452141] R10: 00000000000031dc R11: 7fffffffffffffff R12:
0000000000000001
[147513.459738] R13: ffffffff9f6c9f6d R14: ffff89ec36207e28 R15:
0000000000000087
[147513.466787] FS:  0000000000000000(0000) GS:ffff89ef62500000(0000)
knlGS:0000000000000000
[147513.473205] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[147513.479148] CR2: 00001d110871a000 CR3: 00000001ae210000 CR4:
00000000003607e0
[147513.484452] Call Trace:
[147513.488474]  [<ffffffff9fbc8f1e>] cpuidle_idle_call+0xde/0x230
[147513.491680]  [<ffffffff9f637c8e>] arch_cpu_idle+0xe/0xc0
[147513.494654]  [<ffffffff9f70181a>] cpu_startup_entry+0x14a/0x1e0
[147513.497325]  [<ffffffff9f65a827>] start_secondary+0x1f7/0x270
[147513.499793]  [<ffffffff9f6000d5>] start_cpu+0x5/0x14
[147513.502018] Code: 40 ee b3 ff 4c 89 e6 48 89 df 44 89 f2 49 89 c5
49 8b 47 48 e8 9b f1 dc ff 41 89 c4 e8 23 ee b3 ff 48 89 c7 fb 66 0f
1f 44 00 00 <4c> 29 ef e8 b1 a5 ad ff 48 69 c0 40 42 0f 00 b9 ff ff ff
7f 48

guess i'll wait on sudo and look in the article or on the web for more
information on triggering a crashdump via keystroke.  it's 0723.

0725 the internet sounds like i had it right: hold alt, hit sysrq, hit
the command key.  i'll try again.

it's 0726 and i realized my system has no sysrq key.  i must have hit
some other key, such as scrlk, thinking it was sysrq.  my goodness.
i'll check the proc file and see if it's world writable.


it's 0726 and /proc/sysrq-trigger is writable only by root.  sudo is
still pending.

my toes and fingers have been feeling more cold since my network went
down, possibly due to personal stress since the sun is rising and it
should be getting warmer.  i'll spend some time warming them.

while starting to warm my toes and casually glancing at the page open
on my phone i saw that prtsc may function as sysrq.  i'll try that.
it's 0730

it's 0736.  i took a crashdump; it appeared to work.  after it took
the dump it said it was rebooting, and the screen went blank with a
nonblinking cursor in the upper left.  i used the soft power button to
hard reboot it.
when it came back up, there was no bios, and the screen with the
cursor was still there.  i did the same thing again with disconnecting
all the peripherals, and this time it worked.  i plugged them back in
as it was booting,
and the kernel hung before accepting a decryption passphrase, saying
there was an over current condition on a usb port.  my usb ports are
old and buggy.  i removed the wifi adapter, since it was the focus of
the issues earlier,
and rebooted, and got to the decryption passphrase.  i plugged it back
in and the system booted normally.  i am online again.
it's 0738 and i'm attached to a restored tmux session.

it's 0738 and i'm back in tmux.

here's this:
$ ls -l /var/crash/127.0.0.1-2021-11-06-07\:32\:01/
total 318244
-rw-------. 1 root root 325454858 Nov  6 07:32 vmcore
-rw-r--r--. 1 root root    410893 Nov  6 07:32 vmcore-dmesg.txt
$ sudo sha256sum /var/crash/127.0.0.1-2021-11-06-07\:32\:01/*
[ed, meant to remove this:]sha256sum:
'/var/crash/127.0.0.1-2021-11-06-07:32:01/vmcore': Permission denied
d7133f7f0056961de878f4465c22767cb14f2906d2730be758074526f5ceaa87
/var/crash/127.0.0.1-2021-11-06-07:32:01/vmcore
8187db7cd0e16dec349cd6b552210294861bc003759031a0b30d05ad974a12b3
/var/crash/127.0.0.1-2021-11-06-07:32:01/vmcore-dmesg.txt

i'll boot the gui back up and get this document back into an email so
things make more sense.  that'll also help send the vmcore to redhat
or whatnot.  it's 0740

it's 0742 and i'm working on warming my extremities.  not in email yet.

when i got paranoid and delusional i got chilblains for the first
time.  i was scared and didn't spend enough time rewarming my feet
when outdoors, because i was scared.
i remember repeating to myself when that happened -- or maybe i don't,
all of a sudden, who knows.  anyway: the thing i used to do, was when
my feet were cold (like getting numb), i
would stop whatever i was doing, and just spend time warming them --
until they were completely warm and comfortable.

doing that for a couple years, i ended up with my feet staying warm on
their own.  so it was a decreasing curve, roughly.

i don't remember how i used to warm my feet, though.  but little bits
come back to me here and there.  i think part of it might have been
taking off my socks and holding my bare feet;
i think that might have been part of it.  maybe i'll try that on one
foot now.  i can put the socks from that foot, onto the other one.

it's 0754 and my inhibitions are engaging my foot warming behavior,
making it severely harder to sustain and develop.  i guess i'll kind
of take it adhoc.

things are funny and i have a mitten on my foot.  it's 0755.  my feet
feel okayish, temporarily.

0758.  back in an email.  i'll send the email, and also plan to reboot
the bitcoin clients and zstd processes if reasonable.  my tmux session
wasn't updated because i took a crashdump before closing tmux.