Re: 回复: 回复: [agl-dev-community] About BeagleBone disk space

[Jan-Simon Möller <dl9pf@xxxxxx>]

Hi !

Please check your conf/local.conf :

#>>>>>>>>>

######################
# AGL USER VARIABLES #
######################

# Add additional IMAGE_FSTYPES here on-top of the default ext4.xz and tar.xz
#AGL_EXTRA_IMAGE_FSTYPES = ""

# Additional free disk space created in the image in Kbytes.
#IMAGE_ROOTFS_EXTRA_SPACE="524288"


#<<<<<<<<<

So just adapt IMAGE_ROOTFS_EXTRA_SPACE in your build. 


Long explanation:

The default free space on the images in bitbake is defined as e.g. 5% extra 
space. So the free space in an image depends on the actual image size.
The bbe is building a telematics image which is smaller than the other boards. 
Thus the percentage results in less free space here.

Now what can we do by default: 
* Submit a patch that sets IMAGE_ROOTFS_EXTRA_SPACE:aglci to the BBE config 
fragments. This would only take care of BBE.
or
* Add an OVERRIDE named agltest and submit a patch that sets 
IMAGE_ROOTFS_EXTRA_SPACE:agltest to the agl-test fragments. This would take 
care of all boards.

Latter makes sense as you can quantify your needs. 

Best,
Jan-Simon


Am Mittwoch, 22. März 2023, 10:03:17 CET schrieb Yan Xin Kuan:
> ---
> Where do the 1G and 350MB numbers come from?  Is it that there are testcases
> that write out that much data?
 
> A. That is roughly estimated as upper limit , it is for the future, current
> disk occupation is not that big.
 
> 350MB is the disk occupation raised by testcase output, 1G is installation
> occupation not output data.
 
> ---
> 
> If so, then those tests will need to be made platform specific, I guess.
> 
> A. Like you have said, we are taking this under consideration, so we want to
> make sure that whether BBE is very important for AGL as a long time support
> board, would you please confirm it.
 
> ---
> 
> There may also need to be a discussion on when these tests get run, since if
> they're tied to CI it starts to seem like it could shorten the life of the
> SD cards in some of the platforms. I'd also be concerned about how long
> they'll take if they're tied to CI.
 
> A. That is surely not good, we will set a topic for that later on.
> 
> ---
> -----邮件原件-----
> 发件人: Scott Murray <scott.murray@xxxxxxxxxxxx> 
> 发送时间: 2023年3月22日 14:31
> 收件人: Yan, Xinkuan/晏 新宽 <yanxk.fnst@xxxxxxxxxxx>
> 抄送: agl-dev-community@xxxxxxxxxxxxxxxxxxxxxxxxx; FNST fnstml-fujitsuten
> <fnstml-fujitsuten@xxxxxxxxxxx>
 主题: Re: 回复:  About
> BeagleBone disk space
> 
> On Tue, 21 Mar 2023, yanxk.fnst@xxxxxxxxxxx wrote:
> 
> 
> > Well, the thing is we are checking the disk space for bbe to see if it
> > would meet a situation that is running out of space due to executing
> > "agl-test".
>
> >
> >
> > We have made an estimate of the space required for the final agl-test,
> > which showed that we need:
>
> >
> >
> > 1G for "/usr/" which is mounted on "/"   →   bbe had 710MB left available,
> > so, not enough.
>
> >
> >
> > 350MB for "/var/run" which is mounted on "/run"   →   bbe had 191MB left
> > available, so, not enough either.
>
> >
> >
> > Other boards like H3, rpi4, the QEMUs all look fine.
> >
> >
> >
> > We got data by using agl lava for the disk space info.
> >
> >
> >
> > So, this is it. The BBE would not successfully finish up the agl-test due
> > to in lack of space.
>
> >
> >
> > That is the problem.
> 
> 
> Where do the 1G and 350MB numbers come from?  Is it that there are testcases
> that write out that much data?  If so, then those tests will need to be
> made platform specific, I guess.  There may also need to be a discussion on
> when these tests get run, since if they're tied to CI it starts to seem
> like it could shorten the life of the SD cards in some of the platforms. 
> I'd also be concerned about how long they'll take if they're tied to CI.
 
> Scott
> 
> 
> 
> 
> 






-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#10474): https://lists.automotivelinux.org/g/agl-dev-community/message/10474
Mute This Topic: https://lists.automotivelinux.org/mt/97774096/2167316
Group Owner: agl-dev-community+owner@xxxxxxxxxxxxxxxxxxxxxxxxx
Unsubscribe: https://lists.automotivelinux.org/g/agl-dev-community/leave/4543822/2167316/883735764/xyzzy [list-automotive-discussions82@xxxxxxxxxxx]
-=-=-=-=-=-=-=-=-=-=-=-