DataSourceSmartOS should default to ext4
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
| cloud-init |
Medium
|
Unassigned | ||
| cloud-init (Ubuntu) |
Medium
|
Mike Gerdts |
Bug Description
As reported in Joyent bug IMAGE-1103, ext3 doesn't work with large ephemeral drives.
2018-03-30 23:36:26,529 - util.py[WARNING]: Failed during filesystem operation
Failed to exec of '['/sbin/
Unexpected error while running command.
Command: ['/sbin/mkfs.ext3', '/dev/vdb', '-L', 'ephemeral0', '-F']
Exit code: 1
Reason: -
Stdout:
Stderr: mke2fs 1.42.13 (17-May-2015)
mkfs.ext3: Size of device (0x100000000 blocks) /dev/vdb too big to be expressed
in 32 bits using a blocksize of 4096.
The default should switch to ext4. It so happens that this matches the default in DataSourceAzure.py.
Related branches
- Server Team CI bot: Approve (continuous-integration) on 2018-04-18
- cloud-init Commiters: Pending requested 2018-04-18
-
Diff: 1818 lines (+897/-236)35 files modifiedMANIFEST.in (+1/-0)
bash_completion/cloud-init (+77/-0)
cloudinit/analyze/__main__.py (+1/-1)
cloudinit/config/cc_apt_configure.py (+1/-1)
cloudinit/config/cc_disable_ec2_metadata.py (+12/-2)
cloudinit/config/cc_power_state_change.py (+1/-1)
cloudinit/config/cc_rsyslog.py (+2/-2)
cloudinit/config/tests/test_disable_ec2_metadata.py (+50/-0)
cloudinit/distros/freebsd.py (+3/-3)
cloudinit/net/network_state.py (+5/-6)
cloudinit/netinfo.py (+273/-72)
cloudinit/sources/DataSourceSmartOS.py (+103/-16)
cloudinit/tests/helpers.py (+14/-26)
cloudinit/tests/test_netinfo.py (+101/-85)
cloudinit/util.py (+3/-3)
debian/changelog (+13/-0)
doc/examples/cloud-config-disk-setup.txt (+2/-2)
packages/redhat/cloud-init.spec.in (+1/-0)
packages/suse/cloud-init.spec.in (+1/-0)
setup.py (+1/-0)
tests/cloud_tests/testcases/base.py (+1/-1)
tests/data/netinfo/netdev-formatted-output (+10/-0)
tests/data/netinfo/new-ifconfig-output (+18/-0)
tests/data/netinfo/old-ifconfig-output (+18/-0)
tests/data/netinfo/route-formatted-output (+22/-0)
tests/data/netinfo/sample-ipaddrshow-output (+13/-0)
tests/data/netinfo/sample-iproute-output-v4 (+3/-0)
tests/data/netinfo/sample-iproute-output-v6 (+11/-0)
tests/data/netinfo/sample-route-output-v4 (+5/-0)
tests/data/netinfo/sample-route-output-v6 (+13/-0)
tests/unittests/test_datasource/test_smartos.py (+102/-1)
tests/unittests/test_filters/test_launch_index.py (+5/-5)
tests/unittests/test_merging.py (+1/-1)
tests/unittests/test_runs/test_merge_run.py (+1/-1)
tests/unittests/test_util.py (+9/-7)
- Server Team CI bot: Approve (continuous-integration) on 2018-04-17
- Scott Moser: Approve on 2018-04-17
-
Diff: 35 lines (+3/-3)2 files modifiedcloudinit/sources/DataSourceSmartOS.py (+1/-1)
doc/examples/cloud-config-disk-setup.txt (+2/-2)
Changed in cloud-init (Ubuntu): | |
assignee: | nobody → Mike Gerdts (mgerdts) |
Scott Moser (smoser) wrote : | #1 |
Scott Moser (smoser) wrote : | #2 |
You probably need to change table_type mbr to gpt too if the point is big disks. In fact it might just be the mbr table that is limiting you.
I think we've been avoiding this issue so far because the file system is
being created without partitioning the disk (layout: False). With "layout:
False", I'm guessing that whether gpt or mbr is specified makes no
difference.
Mike
On Thu, Apr 12, 2018 at 8:41 PM, Scott Moser <email address hidden>
wrote:
> You probably need to change table_type mbr to gpt too if the point is
> big disks. In fact it might just be the mbr table that is limiting you.
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https:/
>
> Title:
> DataSourceSmartOS should default to ext4
>
> Status in cloud-init package in Ubuntu:
> New
>
> Bug description:
> As reported in Joyent bug IMAGE-1103, ext3 doesn't work with large
> ephemeral drives.
>
> 2018-03-30 23:36:26,529 - util.py[WARNING]: Failed during filesystem
> operation
> Failed to exec of '['/sbin/
> '-F']':
> Unexpected error while running command.
> Command: ['/sbin/mkfs.ext3', '/dev/vdb', '-L', 'ephemeral0', '-F']
> Exit code: 1
> Reason: -
> Stdout:
> Stderr: mke2fs 1.42.13 (17-May-2015)
> mkfs.ext3: Size of device (0x100000000 blocks) /dev/vdb too big
> to be expressed
> in 32 bits using a blocksize of 4096.
>
> The default should switch to ext4. It so happens that this matches
> the default in DataSourceAzure.py.
>
> To manage notifications about this bug go to:
> https:/
> bug/1763511/
>
Scott Moser (smoser) wrote : | #4 |
ah.yeah, you're probably right.
On Fri, Apr 13, 2018 at 2:02 AM, Mike Gerdts <email address hidden>
wrote:
> I think we've been avoiding this issue so far because the file system is
> being created without partitioning the disk (layout: False). With "layout:
> False", I'm guessing that whether gpt or mbr is specified makes no
> difference.
>
> Mike
>
> On Thu, Apr 12, 2018 at 8:41 PM, Scott Moser <email address hidden>
> wrote:
>
> > You probably need to change table_type mbr to gpt too if the point is
> > big disks. In fact it might just be the mbr table that is limiting you.
> >
> > --
> > You received this bug notification because you are subscribed to the bug
> > report.
> > https:/
> >
> > Title:
> > DataSourceSmartOS should default to ext4
> >
> > Status in cloud-init package in Ubuntu:
> > New
> >
> > Bug description:
> > As reported in Joyent bug IMAGE-1103, ext3 doesn't work with large
> > ephemeral drives.
> >
> > 2018-03-30 23:36:26,529 - util.py[WARNING]: Failed during filesystem
> > operation
> > Failed to exec of '['/sbin/
> > '-F']':
> > Unexpected error while running command.
> > Command: ['/sbin/mkfs.ext3', '/dev/vdb', '-L', 'ephemeral0', '-F']
> > Exit code: 1
> > Reason: -
> > Stdout:
> > Stderr: mke2fs 1.42.13 (17-May-2015)
> > mkfs.ext3: Size of device (0x100000000 blocks) /dev/vdb too big
> > to be expressed
> > in 32 bits using a blocksize of 4096.
> >
> > The default should switch to ext4. It so happens that this matches
> > the default in DataSourceAzure.py.
> >
> > To manage notifications about this bug go to:
> > https:/
> > bug/1763511/
> >
>
> --
> You received this bug notification because you are subscribed to cloud-
> init in Ubuntu.
> https:/
>
> Title:
> DataSourceSmartOS should default to ext4
>
> To manage notifications about this bug go to:
> https:/
> bug/1763511/
>
Changed in cloud-init (Ubuntu): | |
status: | New → In Progress |
importance: | Undecided → Medium |
Scott Moser (smoser) wrote : | #5 |
An upstream commit landed for this bug.
To view that commit see the following URL:
https:/
Launchpad Janitor (janitor) wrote : | #6 |
This bug was fixed in the package cloud-init - 18.2-14-
---------------
cloud-init (18.2-14-
* New upstream snapshot.
- net: Depend on iproute2's ip instead of net-tools ifconfig or route
- DataSourceSmartOS: fix hang when metadata service is down
[Mike Gerdts] (LP: #1667735)
- DataSourceSmartOS: change default fs on ephemeral disk from ext3 to
ext4. [Mike Gerdts] (LP: #1763511)
- pycodestyle: Fix invalid escape sequences in string literals.
- Implement bash completion script for cloud-init command line
-- Chad Smith <email address hidden> Wed, 18 Apr 2018 15:25:53 -0600
Changed in cloud-init (Ubuntu): | |
status: | In Progress → Fix Released |
This bug is believed to be fixed in cloud-init in version 18.3. If this is still a problem for you, please make a comment and set the state back to New
Thank you.
Changed in cloud-init: | |
status: | New → Fix Committed |
importance: | Undecided → Medium |
status: | Fix Committed → Fix Released |
fwiw, I think you could send vendor-data to with cloud-config to set the fs_setup like you wanted (assuming 'you' here are the cloud platform).
The point of vendor data is so that the vendor could make changes like this without changing cloud-init.
But the 1 byte change is also acceptable.
(ext3 -> ext4).