Happy new year!
I've been running Tumbleweed on a laptop since last August. The encrypted LVM partition on the SSD is split into two logical volumes, root and home. Until I upgraded lvm2 two days ago from 2.02.152 to 2.02.165, the system would boot normally. Since the upgrade, the system starts the emergency shell after attempting to mount home fails (displaying "A start job is running for dev-macbook-home.device" until the timer expires). If I type "vgchange -ay" in the shell and exit, boot resumes normally.
As I was checking the snapshot diff before and after the upgrade of lvm2, I noticed that lvm.conf includes extra settings about auto_activation_volume_list etc. Does anyone have any idea how to make lvm behave normally again?
On Sunday, 1 January 2017 12:57:45 EET auxsvr wrote:
I've been running Tumbleweed on a laptop since last August. The encrypted LVM partition on the SSD is split into two logical volumes, root and home. Until I upgraded lvm2 two days ago from 2.02.152 to 2.02.165, the system would boot normally. Since the upgrade, the system starts the emergency shell after attempting to mount home fails (displaying "A start job is running for dev-macbook-home.device" until the timer expires). If I type "vgchange -ay" in the shell and exit, boot resumes normally.
Apparently, the cause was that I hadn't upgraded liblvm2app2_2 and liblvm2cmd2_02 along with lvm2, which caused segfaults when calling lvm2-activation-generator. Shouldn't RPM dependencies handle this case?
On Sunday 2017-01-01 19:42, auxsvr wrote:
On Sunday, 1 January 2017 12:57:45 EET auxsvr wrote:
I've been running Tumbleweed on a laptop since last August. The encrypted LVM partition on the SSD is split into two logical volumes, root and home. Until I upgraded lvm2 two days ago from 2.02.152 to 2.02.165, the system would boot normally. Since the upgrade, the system starts the emergency shell after attempting to mount home fails (displaying "A start job is running for dev-macbook-home.device" until the timer expires). If I type "vgchange -ay" in the shell and exit, boot resumes normally.
Apparently, the cause was that I hadn't upgraded liblvm2app2_2 and liblvm2cmd2_02 along with lvm2, which caused segfaults when calling lvm2-activation-generator. Shouldn't RPM dependencies handle this case?
RPM did what it could, given the information that upstream and the package put into it. Possible causes: - upstream did not specify enough version constraints through symbol versions - upstream did not specify enough constraints through documentation - if upstream did, then the packager did not transplant it into the .spec
-
auxsvr -
Jan Engelhardt