Updating core while it is running causes either SIGSEGV, module not found error, or 100% CPU

[alessiodf]

Updating core while it is running causes either SIGSEGV, module not found error, or 100% CPU

[crnazmija]

Can confirm, it was case number 3 for me:
Process with pid 1316225 still alive after 30000ms, sending it SIGKILL now...

[arktoshidotcom]

Confirmed.
I got the same issue @ajsVoren

[alessiodf]

The issue is caused by the rebuild of better-sqlite3 during the reinstall/update process, not bcrypto as I originally thought.

[sebastijankuzner]

I just want to mention that the next update will probably still fail if you try to update core, because there is still old code executing update. You can prevent fails if you run:

rm -f "yarn global dir"/node_modules/better-sqlite3/build/Release/better_sqlite3.node

before ark update.

Further updates after the next release (running update from 3.1.2 or 3.2.0) will use the latest code, so this line will no longer be needed.

[alessiodf]

I just want to mention that the next update will probably still fail if you try to update core, because there is still old code executing update. You can prevent fails if you run:

rm -f "yarn global dir"/node_modules/better-sqlite3/build/Release/better_sqlite3.node

before ark update.

Further updates after the next release (running update from 3.1.2 or 3.2.0) will use the latest code, so this line will no longer be needed.

Almost!

The line should be

rm -f "`yarn global dir`"/node_modules/better-sqlite3/build/Release/better_sqlite3.node

(Backticks inside the quotes got mangled by Markdown in your message)