Summary: | reboot hang with 'All buffers synced' with FreeBSD13 on ZFS root | ||||||
---|---|---|---|---|---|---|---|
Product: | Base System | Reporter: | Marek <marek> | ||||
Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> | ||||
Status: | Open --- | ||||||
Severity: | Affects Many People | CC: | grahamperrin, nick, sobomax, tschweikle | ||||
Priority: | Normal | ||||||
Version: | 13.1-STABLE | ||||||
Hardware: | Any | ||||||
OS: | Any | ||||||
Attachments: |
|
Description
Marek
2012-09-26 09:50:06 UTC
For bugs matching the following criteria: Status: In Progress Changed: (is less than) 2014-06-01 Reset to default assignee and clear in-progress tags. Mail being skipped FreeBSD 13-Stable does not poweroff after giving "halt -p" or "shutdown -p now". It just hangs after telling uptime. This is for KVM, VMware (Workstation, Player, ESX), VirtualBox, and real Hardware: tested Lenovo T460s. With real hardware you've to press the power button for at least three seconds to power off. Rebooting will take time. After printing uptime you've to wait for up to an hour to see the system reboot. And sometimes it just does not reboot after all. Created attachment 243654 [details]
Screenshot of the problem as it happens with FreeBSD 13.1 kernel.
We are seeing similar issue on our ZFS-enabled AWS instances. Occasionally it stucks on reboot with the "spaw_pager_getswapspace(32): failed" right after "All buffers synced." message (see attached screenshot). In the absence of the real fix, I suggest we maybe add a flag that would panic() the system if swp_pager_getswapspace() fails after "All buffers synced" point. If nothing else, that would give us traceback of where the memory allocation is requested and then actually reboot the machine instead of having it to stuck forever? |