remove kpartx workaround

[adrelanos]

  kpartx -d "${ORIG_TARGET}" >/dev/null
  # Workaround for a bug in kpartx which doesn't clean up properly,
  # see Debian Bug #891077 and Github-PR grml/grml-debootstrap#112
  if dmsetup ls | grep -q "^${LOOP_PART} "; then
    kpartx -d "/dev/${LOOP_DISK}" >/dev/null
  fi

• https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=891077
• https://github.com/grml/grml-debootstrap/pull/112

Should be fixed? Can this workaround be removed?

[jkirk]

Ah, thanks for bringing this up.

I don't have the steps to reproduce for the problem with kpartx anymore (we worked on this 7 years ago at the Debconf18 in Taiwan. Brings back memories. 😊 ).

Did you test grml-debootstrap without that "workaround"? Did it work reliably? But now that #891077 is fixed, I think we should be safe to remove the workaround.

/cc @jkirk @zeha

[zeha]

Ack, somebody please test and confirm this can go.

[adrelanos]

The mentioned Debian bug report states fixed indeed. However, there seems to be a regression.

https://github.com/grml/grml-debootstrap/issues/345 seems like a valid report.

The best outcome would be:

• short term: fix https://github.com/grml/grml-debootstrap/issues/345
• long term: report regression issue upstream.

(This bug report was from theoretic review of the source code. Also because I didn't experience the issue. I thought it's a good idea to ask if anyone knows more about it. A few days later, I learned that however this is still an issue in some situations, will elaborate in #345.)

[mika]

Ah, interesting, thanks for looking into it, @adrelanos! Did the dmsetup output or alike change, or why wasn't this a problem in the past? 🤔