New issue
Advanced search Search tips

Issue 919190 link

Starred by 2 users
Status: Assigned
Owner:
dverkamp@chromium.org
Cc:
smbar...@chromium.org
dverkamp@chromium.org
dgreid@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: Bug



Sign in to add a comment

grunt: StartTermina fails with "failed to mount stateful"

Project Member Reported by smbar...@chromium.org, Jan 4 
vm.CrostiniStartEverything is failing occasionally on grunt to mount the stateful disk image: https://stainless.corp.google.com/search?status=FAIL&status=ERROR&status=ABORT&test=%5Etast%5C.vm%5C.&reason=failed+to+mount+stateful&exclude_non_release=true&exclude_cts=true&view=list&first_date=2018-12-08&last_date=2019-01-04

It looks like the virtio-pci device is missing in those logs. There should be a device at 0000:00:02.0

[    4.341564] virtio-pci 0000:00:01.0: enabling device (0000 -> 0002)
[    4.352534] virtio-pci 0000:00:01.0: PCI->APIC IRQ transform: INT A -> IRQ 5
[    4.358976] virtio-pci 0000:00:03.0: enabling device (0000 -> 0002)
[    4.361938] virtio-pci 0000:00:03.0: PCI->APIC IRQ transform: INT C -> IRQ 7
[    4.368114] virtio-pci 0000:00:04.0: enabling device (0000 -> 0002)
[    4.371065] virtio-pci 0000:00:04.0: PCI->APIC IRQ transform: INT D -> IRQ 8
[    4.377288] virtio-pci 0000:00:05.0: enabling device (0000 -> 0002)
[    4.380266] virtio-pci 0000:00:05.0: PCI->APIC IRQ transform: INT A -> IRQ 9
[    4.386453] virtio-pci 0000:00:06.0: enabling device (0000 -> 0002)
[    4.389432] virtio-pci 0000:00:06.0: PCI->APIC IRQ transform: INT B -> IRQ 10
[    4.396130] virtio-pci 0000:00:07.0: enabling device (0000 -> 0002)
[    4.399377] virtio-pci 0000:00:07.0: PCI->APIC IRQ transform: INT C -> IRQ 11

Comment 1 by dverkamp@chromium.org, Jan 4 

It looks like the virtio device child process is exiting for some reason early in startup - I see this earlier in the logs in all 3 cases:

2019-01-04T00:57:09.237402-08:00 ERR localhos[18447]: crosvm[4]: [devices/src/proxy.rs:178] failed read from child device process: Recv(Error(11))

(In fact, one of the failures is actually missing both PCI devices 1 and 2, and it has the same error repeated twice, which reinforces my guess that this is related to the problem.)

That said, I'm not sure if the device process is actually gone, or if it's something less severe - errno 11 is EAGAIN, and we do have a timeout set on the ProxyDevice socket, so we could just be timing out early (the timeout is 2000 ms, though, so I don't know if that's likely).

Comment 2 by dgreid@google.com, Jan 10

Labels: -Pri-2 M-73 Pri-1
Owner: dverkamp@chromium.org
Status: Assigned (was: Untriaged)
►

Sign in to add a comment