Running Mir-Server snap, gives error! [Unknown command line options: --vt 1]

[Janne Hannila]

I don't know, if this is right place!
Trying to run mir-server gives error:

Dec 8 02:59:06 localhost snap[4768]: Unknown command line options: --vt 1
Dec 8 02:59:06 localhost systemd[1]: snap.mir-server.mir-server.service: Main process exited, code=exited, status=1/FAILURE
Dec 8 02:59:06 localhost systemd[1]: snap.mir-server.mir-server.service: Unit entered failed state.
Dec 8 02:59:06 localhost systemd[1]: snap.mir-server.mir-server.service: Failed with result 'exit-code'.
Dec 8 02:59:06 localhost systemd[1]: snap.mir-server.mir-server.service: Service hold-off time over, scheduling restart.

Running on Raspberry pi 2, Ubuntu snappy core!

runned lsb_release -rd

Description: Ubuntu 16.04.1 LTS
Release: 16.04

[Daniel van Vugt (vanvugt)]

The error 'Unknown command line options: --vt 1' means you don't have (or aren't using?) the mesa-kms graphics driver needed to interpret that option. I don't know if the mir-server snap has correctly provided this.

[Daniel van Vugt (vanvugt)]

Incidentally in future I think the VT option should be server-wide. It's not graphics-driver specific, however some kernels don't have VT support (ChromeOS??) so we need to be aware of that too.

[Janne Hannila (janne-hannila)]

Here before that error lists
Dec 8 03:35:39 localhost snap[2590]: [2016-12-08 03:35:39.561979] mirplatform: Found graphics driver: mir:mesa-kms (version 0.24.1)
Dec 8 03:35:39 localhost snap[2590]: [2016-12-08 03:35:39.562510] mirplatform: Found graphics driver: mir:mesa-x11 (version 0.24.1)

[Daniel van Vugt (vanvugt)]

Thanks for pointing that out. So it's probably not a bug in the snap itself but a Mir bug.

I have seen this happen before in other bug reports (spurious 'Unknown command line options: --vt 1' even when the mesa-kms driver is present), but will need to find those old bug reports...

[Daniel van Vugt (vanvugt)]

Is there another message in the log saying "mirserver: Selected driver: ..."?

[Daniel van Vugt (vanvugt)]

Please let us know what the "Selected driver" log message was. Or better yet, attach the full log output.

A similar error was encountered in bug 1355005 and although that bug is unclear it seems to generally be a case of Mir selecting a driver other than mesa-kms, and if that happens then such other drivers will reject the --vt 1 option.

So if you Raspberry Pi is selecting the mesa-kms driver then yes we have a Mir bug. If it's not and is somehow selecting the Android driver instead (and if that actually works) only then is it a bug in the snap (should not be passing --vt 1).

[Janne Hannila (janne-hannila)]

Hi,

As you see, it tries to load it few times.
This is complete syslog for it:

Dec 8 08:11:16 localhost systemd[1]: Started Service for snap application mir-server.mir-server.
Dec 8 08:11:17 localhost snap[1608]: Mir server snap started
Dec 8 08:11:18 localhost snap[1608]: [2016-12-08 08:11:18.206640] mirplatform: Found graphics driver: mir:mesa-kms (version 0.24.1)
Dec 8 08:11:18 localhost snap[1608]: [2016-12-08 08:11:18.275107] mirplatform: Found graphics driver: mir:mesa-x11 (version 0.24.1)
Dec 8 08:11:18 localhost snap[1608]: Unknown command line options: --vt 1
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Main process exited, code=exited, status=1/FAILURE
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Unit entered failed state.
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Failed with result 'exit-code'.
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Service hold-off time over, scheduling restart.
Dec 8 08:11:18 localhost systemd[1]: Stopped Service for snap application mir-server.mir-server.
Dec 8 08:11:18 localhost systemd[1]: Started Service for snap application mir-server.mir-server.
Dec 8 08:11:18 localhost snap[1621]: Mir server snap started
Dec 8 08:11:18 localhost snap[1621]: [2016-12-08 08:11:18.759812] mirplatform: Found graphics driver: mir:mesa-kms (version 0.24.1)
Dec 8 08:11:18 localhost snap[1621]: [2016-12-08 08:11:18.760352] mirplatform: Found graphics driver: mir:mesa-x11 (version 0.24.1)
Dec 8 08:11:18 localhost snap[1621]: Unknown command line options: --vt 1
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Main process exited, code=exited, status=1/FAILURE
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Unit entered failed state.
Dec 8 08:11:18 localhost systemd[1]: snap.mir-server.mir-server.service: Failed with result 'exit-code'.
Dec 8 08:11:19 localhost systemd[1]: snap.mir-server.mir-server.service: Service hold-off time over, scheduling restart.
Dec 8 08:11:19 localhost systemd[1]: Stopped Service for snap application mir-server.mir-server.
Dec 8 08:11:19 localhost systemd[1]: Started Service for snap application mir-server.mir-server.
Dec 8 08:11:19 localhost snap[1633]: Mir server snap started
Dec 8 08:11:19 localhost snap[1633]: [2016-12-08 08:11:19.199499] mirplatform: Found graphics driver: mir:mesa-kms (version 0.24.1)
Dec 8 08:11:19 localhost snap[1633]: [2016-12-08 08:11:19.200065] mirplatform: Found graphics driver: mir:mesa-x11 (version 0.24.1)
Dec 8 08:11:19 localhost snap[1633]: Unknown command line options: --vt 1
Dec 8 08:11:19 localhost systemd[1]: snap.mir-server.mir-server.service: Main process exited, code=exited, status=1/FAILURE
Dec 8 08:11:19 localhost systemd[1]: snap.mir-server.mir-server.service: Unit entered failed state.
Dec 8 08:11:19 localhost systemd[1]: snap.mir-server.mir-server.service: Failed with result 'exit-code'.
Dec 8 08:11:19 localhost systemd[1]: snap.mir-server.mir-server.service: Service hold-off time over, scheduling restart.
Dec 8 08:11:19 localhost systemd[1]: Stopped Service for ...

[Janne Hannila (janne-hannila)]

So i can't find "mirserver: Selected driver"

[Daniel van Vugt (vanvugt)]

OK in that case I can't tell what's going wrong. We'll need to look at (or ask the creator of) the snap itself.

The bug can probably stay incomplete because there's not yet sufficient evidence that this is a bug in Mir at all. Perhaps just a bug in the snap definition.

[kevin gunn (kgunn72)]

sorry this happened off my timezone, but I can share with you that the RPi2/3 graphics are not quite ready yet. There is a patch set requiring integration into the ubuntu kernel that is maintained for RPi. There is work ongoing in the kernel team now to get these patches integrated, validated and released. If i can find the related bug, I will post the link.
We can eventually mark this invalid....but let's leave incomplete until we can find the kernel bug link.

[Janne Hannila (janne-hannila)]

Same problem in mir-kiosk from canonical!

[Daniel van Vugt (vanvugt)]

Even if we don't have the kernel support in place yet I think Mir could still do better.

What the user sees is:
Unknown command line options: --vt 1

What that means is:
A non-mesa-kms graphics driver was chosen (or none at all? can't tell). And that's not even a bug at all if some working driver has been found. It's presently a bug in the snap to pass a driver-specific option (--vt), but I would also argue that should not be true and the snap should be allowed to (along with LightDM). VT support should be generally available and not buried in just one graphics driver. Then the above error will never happen again (we've been hitting it for years now).

[Alan Griffiths (alan-griffiths)]

There are two problems being discussed:

/1/ mir doesn't work on the current Ubuntu snappy core kernel for Raspberry pi 2; and,
/2/ Mir does not give good diagnostic information from our attempt at automaticity deciding on the right graphics stack.

IMO systems that try a series of options (in this case graphics modules) and only report problems when the last one fails are a PITA, but that has been discussed many times. I don't see an existing bug for this. Maybe we should open one and leave this report for the former problem?

[Daniel van Vugt (vanvugt)]

Incomplete I think is still the correct status here.

Although this bug report has reminded us of some things we don't like about Mir, it hasn't yet shown that there is an underlying problem in Mir causing this particular bug.

Unfortunately the log we're seeing from the snap looks kind of incomplete so it's hard to tell. So that's a second issue really where we want to log a bug against the snap definition rather than Mir. Where do we log bugs against the snap?

[Launchpad Janitor (janitor)]

[Expired for mir (Ubuntu) because there has been no activity for 60 days.]

[Launchpad Janitor (janitor)]

[Expired for Mir because there has been no activity for 60 days.]