| View previous topic :: View next topic |
| Author |
Message |
klammerj
n00b
Joined: 19 Jan 2023
Posts: 14
|
 Posted: Thu May 09, 2024 6:58 pm Post subject: How do I debug those openrc hangs on shutdown? |
 |
|
Good evening,
Typically there's about a 10percent chance the box locks up.
often it's after or at localmount or sysklogd, so the kernel logs won't be much help.
I'd need console output. However, occasional modesets during shutdown make it difficult to impossible
to guess what services are still running and what are not.
I've done a fair bit of guessing and changing the need <whatever> clauses.
So far without much effect.
I seem to have version 0.48 of openrc installed. |
|
| Back to top |
|
 |
ian.au
Guru
Joined: 07 Apr 2011
Posts: 595
Location: Australia
|
| Posted: Fri May 10, 2024 1:25 am Post subject: |
|
|
klammerj,
openrc stable is at 0.54 and 0.48 was replaced in January here, so I assume you're holding off updating your box? Why is that?
There isn't any useful info beyond that in your post, so everybody would be playing the same guessing game as you, your problem could be almost anything. If you've set your kernel for parallel execution the observed location of the hang might be quite misleading.
| Code: | gw-01 ~ # grep -i parallel /usr/src/linux/.config
CONFIG_HOTPLUG_PARALLEL=y |
I'd say the likely culprit is either network or ACPI configuration - if it's the latter there should be some failed services in dmesg. You can try | Code: | | gw-01 ~ # dmesg | grep -i error |
and see if that helps.
If network, it will depend on your network management setup. If you're running any sort of server or connecting to remote fs via NFS or similar you should mention those.
You should try posting a description of your hardware, lspci etc, emerge --info, dmesg, network management, any remote fs services and kernel configs here for someone to take a look at if you really want help with this. |
|
| Back to top |
|
|
klammerj
n00b
Joined: 19 Jan 2023
Posts: 14
|
| Posted: Fri May 10, 2024 8:18 am Post subject: |
|
|
| ian.au wrote: | klammerj,
openrc stable is at 0.54 and 0.48 was replaced in January here, so I assume you're holding off updating your box? Why is that?
|
Last time I did a larger emerge it took a week or so to fix everything that got hosed in the process.
I'll update that pkg and see what happens...
| ian.au wrote: |
There isn't any useful info beyond that in your post, so everybody would be playing the same guessing game as you, your problem could be almost anything. If you've set your kernel for parallel execution the observed location of the hang might be quite misleading.
| Code: | gw-01 ~ # grep -i parallel /usr/src/linux/.config
CONFIG_HOTPLUG_PARALLEL=y |
|
I'm using distribution kernel 6.1.60-gentoo-dist
grep -i parallel /usr/src/linux/.config
# Raw/parallel NAND flash controllers
# CONFIG_AD7606_IFACE_PARALLEL is not set
| ian.au wrote: |
I'd say the likely culprit is either network or ACPI configuration - if it's the latter there should be some failed services in dmesg. You can try | Code: | | gw-01 ~ # dmesg | grep -i error |
and see if that helps.
|
dmesg | grep -i error
[ 0.268797] acpi PNP0A08:00: _OSC: platform retains control of PCIe features (AE_ERROR)
[ 1.013014] RAS: Correctable Errors collector initialized.
[ 22.396100] dracut: Mounting /dev/sda2 with -o defaults,noatime,lazytime,noiversion,inode_readahead_blks=2,delalloc,errors=remount-ro
[ 45.335372] platform regulatory.0: Direct firmware load for regulatory.db failed with error -2
| ian.au wrote: |
If network, it will depend on your network management setup. If you're running any sort of server or connecting to remote fs via NFS or similar you should mention those.
You should try posting a description of your hardware, lspci etc, emerge --info, dmesg, network management, any remote fs services and kernel configs here for someone to take a look at if you really want help with this. |
Yes, there's network shares involved and kerberos and gssd.. but
I'd rather not divulge too much about my network setup.
I know it sez `noob' there on the left, but that's just coz I rarely post to fora. I've been using this stuff since about 1990..
What I'd like is just some way to print [123689] before every shutdown script to indicate what
services are still running(and a way to assign arbitrary numbers
to the individual service scripts). I'd be done debugging this in an hour,
instead of messing around for months now. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
