Project

General

Profile

Actions
Copy link

Bug #12373

closed

sgdisk hangs on zap command

Added by Chris Holcombe almost 9 years ago. Updated over 8 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
other
Tags:
Backport:
Regression:
No
Severity:
2 - major
Reviewed:
Affected Versions:
v0.94
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

The launchpad bug that contains all the info:
https://bugs.launchpad.net/charms/+source/ceph/+bug/1475247

What we're seeing on several machines now is that the /sbin/sgdisk --zap-all --clear --mbrtogpt -- /dev/sdb is taking upwards of 40-50 minutes to complete. According to the sgdisk website the "--clear" command will fail if the partition on the disk is corrupted in any way. It's possible that this is calling sgdisk to hang and wait in a D state that cannot be killed. I'm going to try patching the python code to do 2 steps instead of one: 1) sgdisk --zap-all 2) sgdisk --clear --mbrtogpt -- /dev/sdb and see if that helps.

Actions
Copy link

Also available in: Atom PDF