On 03/18/2017 05:21 PM, Unman wrote: > On Sat, Mar 18, 2017 at 08:59:29AM -0700, Grzesiek Chodzicki wrote: >> W dniu sobota, 18 marca 2017 15:08:35 UTC+1 użytkownik Unman napisał: >>> On Sat, Mar 18, 2017 at 03:41:49AM -0700, Grzesiek Chodzicki wrote: >>>> I've experienced this issue myself and I saw some people complaining about >>>> it on the mailing list. Sometimes Qubes takes forever to poweroff. >>>> >>>> The issue is caused by the fact that for some unknown reason machines with >>>> PCI devices attached take forever to shutdown and sometimes need to be >>>> manually killed from within VM Manager. After manually killing them, the >>>> system will shutdown normally. >>>> >>>> So, here's the fix: >>>> >>>> In Dom0 run qvm-prefs -s sys-net pci_e820_host false >>>> >>>> Repeat that for sys-usb and any other machine that holds a PCI device. >>>> >>>> Your system should shut down normally after this. >>>> >>>> No, I don't know how or why this works but I know that it solved the issue >>>> on my machine. Explanation welcome. I'll add this to documentation once we >>>> have a confirmation that this is 100% reproducible. >>>> >>> I'd be reluctant to propose this as a tip as is, because it will almost >>> certainly break networking for some users. >>> We specifically set pci_e820 to address user problems. I don't know what >>> the intersection is between those for whom this is a fix and those for >>> whom shutdown is slow/doesn't happen. >>> >>> Have you tried the solution suggested here before of shutting down other >>> vms first, then sys-usb and sys-net? Andrew, I think, posted a simple >>> script that he uses. I suspect that that will have more impact with >>> fewer consequences than this proposal, but user experience will be key >>> here: comments ? >> Yes, I did try that, this is how I found out that on my system it is not >> possible to cleanly shut down VMs that hold PCI devices. If I try shutting >> down sys-net the VM Manager hangs for a few minutes, then the entire system >> stops responding then the machine gets killed, then VM manager throws an >> exception and then system becomes responsive again. After setting >> pci_e820_host to False I can cleanly shutdown sys-net and sys-usb just like >> any other VM. >> > So that sounds very system specific. Why not put together a tip > incorporating Andrew's script ( which will help on almost all > machines), and then yours as a special case to be used where that > doesnt help? That's the point, I don't know whether this is specific to my machine but I don't have another PC to test so I posted this here so other people can try it and see if that alleviates the issue.
-- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/b4834e61-6e58-6a2a-a552-81e9509d007b%40gmail.com. For more options, visit https://groups.google.com/d/optout.
