Date: Tue, 30 May 2017 17:22:11 +0100 From: Pete French <petefrench@ingresso.co.uk> To: bapt@FreeBSD.org, petefrench@ingresso.co.uk Cc: stable@freebsd.org Subject: Re: 11.1-PRERELEASE, waagent crashes attaching reosurce disc in Azure Message-ID: <E1dFjuN-0000l0-0c@dilbert.ingresso.co.uk> In-Reply-To: <e4fee006-a1a0-b25e-3835-ae9f857f171c@ingresso.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
A bit more investigation on this and I think I found what the differece is at least. waagent is using camcontrol periphlist 3:1:0 to find devices, and this has chnaged from placing the discs before the pass devices to afterwards. So on an earlier version I get this: root@joanna-may:/home/webadmin # camcontrol periphlist 3:1:0 da1: generation: 4 index: 1 status: MORE pass1: generation: 4 index: 2 status: LAST and on the box with the issue it does this: root@backup-db:/home/webadmin # camcontrol periphlist 3:1:0 pass1: generation: 4 index: 1 status: MORE da1: generation: 4 index: 2 status: LAST So that box finds 'pass1' as the device when it should be finding 'da1' which leads to the error. I have a vague memory of seeing a commit for this go past too. On the other hand, the oython looks like it shuold handle this, as it appears to try and detect which devices are disks from the output of gpart list, which should not let it use pass1 as if it was a disc. -pete.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E1dFjuN-0000l0-0c>