With micro-osd.sh supporting a 2nd alternate fs we add basic support to
entrypoint.sh to help the tests find and use that alternate fs.
The command line option is flexible, you can supply it with a literal fs
name or, when prefixed with an `@`, it will read the name from a file.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Add a function to bring up a 2nd cephfs file system. Execute it on
ceph versions pacific and later.
While the API exists on nautilus and octopus those versions are already
deprecated in go-ceph. Additionally, the ceph docs for "multifs" state,
"Beginning with the Pacific release, multiple file system support is
stable and ready to use." [1] Therefore we won't bother running our test
automation on anything earlier.
1 - https://docs.ceph.com/en/latest/cephfs/multifs/#multiple-ceph-file-systems
Signed-off-by: John Mulligan <jmulligan@redhat.com>
The linebreaks within the 2 bottom paragraphs look fine in an editor
such as vim, but look odd when viewed as part of the PR on github.
Improve the visuals and make it more readable in the github UI by
removing linebreaks.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Fixes: #815
When updates to the NFS mgr API were recently merged on Ceph main branch
the go-ceph test started failing. This occurred because go-ceph was
unintentionally relying on invalid JSON (an empty string) returned from
the API call for an invalid export pseudo path. On ceph main it now
returns an empty JSON object. This change treats both empty string
and empty JSON object as an error explicitly.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Use the GOARCH variable to define to proper version of go binaries to
download. Without this, the end result image can not run properly on
local host machine.
Signed-off-by: Shachar Sharon <ssharon@redhat.com>
Pass GOARCH as explicit argument when building container image. Use
local host machine to define desired architecture.
Signed-off-by: Shachar Sharon <ssharon@redhat.com>
Following APIs are being added as preview APIs:
FSAdmin.GetMetadata
FSAdmin.SetMetadata
FSAdmin.RemoveMetadata
FSAdmin.ForceRemoveMetadata
FSAdmin.ListMetadata
FSAdmin.GetSnapshotMetadata
FSAdmin.SetSnapshotMetadata
FSAdmin.RemoveSnapshotMetadata
FSAdmin.ForceRemoveSnapshotMetadata
FSAdmin.ListSnapshotMetadata
Signed-off-by: Anoop C S <anoopcs@cryptolab.net>
As of d6bf51115d we've tried to make API
stability less burdensome on contributors by putting in placeholder
values in the status files. However, we haven't used it much yet. This
change executes `make api-fix-versions` in order to convert
the placeholder versions into real version numbers - both for the
real reason of tracking the values and also for a more real world test
of the apiage feature.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
It now reports "No Preview/Deprecated APIs found. All APIs are
considered stable." rather than leaving a blank section.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
When generating the human-readable api-status.md document from the
JSON, the logic used to determine if a package had no Preview/Deprecated
APIs was incorrect. This change fixes and simplifies that condition.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Add a new checklist item reminding contributors to run `make
api-update` to track new APIs. Add two short paragraphs that
cover some of the common things I copy and paste or need to look up
when working with contributors' PRs.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Adds a info and a warning function that currently just wraps print.
In the future these classified print statements could be customized
for output specific to the message type.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Previously, apiage.py just print out a problem count when it found
an unexpected stable api in the new apis file. Now the output displays
something like this:
```
found new unexpected stable apis: API.LeakyBucket
error: 1 problems detected
```
Signed-off-by: John Mulligan <jmulligan@redhat.com>
This is a left over from v0.19.0 release where we missed to remove
preview tag when corresponding API was promoted to stable.
Signed-off-by: Anoop C S <anoopcs@cryptolab.net>
These are left overs from v0.13.0 and v0.17.0 releases where we missed
to remove preview tag when corresponding APIs were promoted to stable.
Signed-off-by: Anoop C S <anoopcs@cryptolab.net>
The v0.16.0 has a "sword" marker noting that quincy was working but not
officially fully supported in v0.16. Now that v0.16 is a few releases in
the past we can simplify that note down to the minimal level of details.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
There was continued flakiness when the tests that enabled the cephfs
mirroring module were enabled. This change adds another 200 ms delay
to hopefully cover the time that might occur as the mgr has a new
module enabled and restart. But unfortunately this is still a bit of
a wild guess.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
This is a manual workflow, you must trigger it from the github ui,
but most everything after that is automated.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
I personally do not like the style of using one .PHONY line as it just
means I end up forgetting to mark things as phony. Change all the api-*
targets to use a .PHONY: per target so its very clear these are all
phony.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Work around an issue reported that due to a large number of items in the
rbd trash the function fails rather than listing the items as desired.
This short term fix simply increases the limit rather than doing
anything long term.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
Turns out the dependabot tool can also indicate what github actions we
are using can be updated. This can be handy in avoiding the warnings
that github throws up all over an action that relies on a deprecated
feature. Use a monthly update just like go modules to avoid being
pestered all the time.
Signed-off-by: John Mulligan <jmulligan@redhat.com>
