Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

percona-cluster 8.0.x sst crashes #380

Open
xmerlin opened this issue Apr 25, 2024 · 3 comments
Open

percona-cluster 8.0.x sst crashes #380

xmerlin opened this issue Apr 25, 2024 · 3 comments
Assignees

Comments

@xmerlin
Copy link

xmerlin commented Apr 25, 2024

After resolving the encoding issue of bug #377, I re-tested Percona Cluster 8.0.x and, although the MySQL client crash is no longer present, the replication and cluster initialization do not work properly. The primary machine set to provide the initial information does not complete the transfer to the joiners correctly, and various MySQL processes are marked as . I tested up to version 8.0.36.

jperkin pushed a commit that referenced this issue May 4, 2024
5.0.18.4, released 2024-01-14
    Don't test on GHC 9.2 or earlier
    Fix up the output of /stats to account for API changes
    #380, move results before left anchors in tab focus order
@xmerlin
Copy link
Author

xmerlin commented May 4, 2024

this patch solves the issue
patch-scripts_wsrep__sst__xtrabackup-v2.diff.gz

@jperkin
Copy link
Collaborator

jperkin commented May 8, 2024

Thanks for looking at this. I think that will require that coreutils be a DEPENDS for the package, as nproc is only available if that package is installed. One drawback of using nproc is that it doesn't account for the actual CPU resources available to the zone, and will simply return the number of CPU cores available on the entire system. A way to do that without having to use an external package would be something like psrinfo | wc -l.

Doing something like mdata-get sdc:cpu_cap and calculating based on that would give you accurate numbers, but is only available to the root user.

I'll have a think...

@jperkin jperkin self-assigned this May 8, 2024
@xmerlin
Copy link
Author

xmerlin commented May 12, 2024

I've simply reused a section of a previous patch get_proc is already there so we can you it also for the decompress task

jperkin pushed a commit that referenced this issue May 17, 2024
0.21.0 (2024-03-10)
-------------------

- Improve documentation. [#483]

- Add a minimum version requirement for ``asdf-wcs-schemas``. [#491]

- Fix ``WCS.__str__`` for instances without transforms. [#489]

0.20.0 (2023-11-29)
-------------------

- Replace ``pkg_resources`` with ``importlib.metadata``. [#478]

- Serialize and deserialize ``pixel_shape`` with asdf. [#480]

0.19.0 (2023-09-15)
-------------------

Bug Fixes
^^^^^^^^^

- Synchronize ``array_shape`` and ``pixel_shape`` attributes of WCS
  objects. [#439]

- Fix failures and warnings with numpy 2.0. [#472]

other
^^^^^

- Remove deprecated old ``bounding_box``. The new implementation is released with
  astropy v 5.3. [#458]

- Refactor ``CoordinateFrame.axis_physical_types``. [#459]

- ``StokesFrame`` uses now ``astropy.coordinates.StokesCoord``. [#452]

- Dropped support for Python 3.8. [#451]

- Fixed a call to ``astropy.coordinates`` in ``wcstools.wcs_from_points``. [#448]

- Code and docstrings clean up. [#460]

- Register all available asdf extension manifests from ``asdf-wcs-schemas``
  except 1.0.0 (which contains duplicate tag versions). [#469]

- Register empty extension for 1.0.0 to avoid warning about a missing
  extension when opening old files. [#475]


0.18.3 (2022-12-23)
-------------------
Bug Fixes
^^^^^^^^^

- Fixed a bug in the estimate of pixel scale in the iterative inverse
  code. [#423]

- Fixed constant term in the polynomial used for SIP fitting.
  Improved stability and accuracy of the SIP fitting code. [#427]


0.18.2 (2022-09-07)
-------------------
Bug Fixes
^^^^^^^^^

- Corrected the reported requested forward SIP accuracy and reported fit
  residuals by ``to_fits_sip()`` and ``to_fits()``. [#413, #419]

- Fixed a bug due to which the check for divergence in ``_fit_2D_poly()`` and
  hence in ``to_fits()`` and ``to_fits_sip()`` was ignored. [#414]

New Features
^^^^^^^^^^^^

0.18.1 (2022-03-15)
-------------------
Bug Fixes
^^^^^^^^^

- Remove references to the ``six`` package. [#402]

0.18.0 (2021-12-22)
-------------------
Bug Fixes
^^^^^^^^^

- Updated code in ``region.py`` with latest improvements and bug fixes
  from ``stsci.skypac.regions.py`` [#382]

- Added support to ``_compute_lon_pole()`` for computation of ``lonpole``
  for all projections from ``astropy.modeling.projections``. This also
  extends support for different projections in ``wcs_from_fiducial()``. [#389]

New Features
^^^^^^^^^^^^

- Enabled ``CompoundBoundingBox`` support for wcs. [#375]

- Moved schemas to standalone package ``asdf-wcs-schemas``.
  Reworked the serialization code to use ASDF converters. [#388]

0.17.1 (2021-11-27)
-------------------

Bug Fixes
^^^^^^^^^

- Fixed a bug with StokesProfile and array types. [#384]


0.17.0 (2021-11-17)
-------------------
Bug Fixes
^^^^^^^^^

- `world_axis_object_components` and `world_axis_object_classes` now ensure
  unique keys in `CompositeFrame` and `CoordinateFrame`. [#356]

- Fix issue where RuntimeWarning is raised when there are NaNs in coordinates
  in angle wrapping code [#367]

- Fix deprecation warning when wcs is initialized with a pipeline [#368]

- Use ``CD`` formalism in ``WCS.to_fits_sip()``. [#380]


New Features
^^^^^^^^^^^^
- ``wcs_from_points`` now includes fitting for the inverse transform. [#349]

- Generalized ``WCS.to_fits_sip`` to be able to create a 2D celestial FITS WCS
  from celestial subspace of the ``WCS``. Also, now `WCS.to_fits_sip``
  supports arbitrary order of output axes. [#357]


API Changes
^^^^^^^^^^^
- Modified interface to ``wcs_from_points`` function to better match analogous function
  in astropy. [#349]

- ``Model._BoundingBox`` was renamed to ``Model.ModelBoundingBox``. [#376, #377]

0.16.1 (2020-12-20)
-------------------
Bug Fixes
^^^^^^^^^
- Fix a regression with ``pixel_to_world`` for output frames with one axis. [#342]

0.16.0 (2020-12-18)
-------------------
New Features
^^^^^^^^^^^^

- Added an option to `to_fits_sip()` to be able to specify the reference
  point (``crpix``) of the FITS WCS. [#337]

- Added support for providing custom range of degrees in ``to_fits_sip``. [#339]

Bug Fixes
^^^^^^^^^

- ``bounding_box`` now works with tuple of ``Quantities``. [#331]

- Fix a formula for estimating ``crpix`` in ``to_fits_sip()`` so that ``crpix``
  is near the center of the bounding box. [#337]

- Allow sub-pixel sampling of the WCS model when computing SIP approximation in
  ``to_fits_sip()``. [#338]

- Fixed a bug in ``to_fits_sip`` due to which ``inv_degree`` was ignored. [#339]
github-actions bot pushed a commit that referenced this issue Oct 21, 2024
# rio 1.2.3

* Fix #453, don't nudge the user to install all suggested packages
* Fix #451, don't nudge the user to report issues about the trust parameter

# rio 1.2.2

* Fix #447 - remove an ancient artefact of Vignette generation, h/t
  Tim Taylor for the help.

# rio 1.2.1

* Roll back the decision to add parquet in the import tier see #455 #315

# rio 1.2.0

* Fix lintr issues #434 (h/t @Bisaloo Hugo Gruson)
* Drop support for R < 4.0.0 see #436
* Add support for parquet in the import tier using `nanoparquet` see
  rio 1.0.1 below.

Bug fixes

* Fix #430 Add back support for `.dat`

# rio 1.1.1

Bug fixes

* Fix #425 for archive formats, the file extension of the input file
  is determined by the compressed file (like prior rio 1.1.0)

# rio 1.1.0

* CRAN release

# rio 1.0.3

* Add `trust` parameter to functions that are used to load various R
  environment formats (`.R`, `.Rds`, `.Rdata`, etc). This parameter is
  defaulted to `TRUE` today to ensure backwards compatibility. A
  deprecation notice warns this will default to `FALSE` in `rio`
  2.0. We are informing users that these data types should only be
  loaded from trusted sources, which should be affirmatively attested
  to.

* Test and fix the compression mechanism: Gzip, Bzip2 are now working as expected.

Bug fixes

- Fix #412, prevent double usage of `which` for archive formats
- Fix #415, both `import_list()` and `export_list()` support tar archives.
- Fix #421, tar export is only supported by R >= 4.0.3.

# rio 1.0.2

* For missing files in `import_list` it gives more informative
  warnings fix #389
* Single-item list of data frames can be exported fix #385
* Move `stringi` to Suggests to reduce compilation time. Add an
  attribution to the internal data to list out all required packages
  #378
* Move `readr` to Imports for `fwf`. `readr` is a dependency of
  `haven` so it does not increase the number of dependencies. Remove
  the original `read.fwf2` which doesn't guess `widths`. Keep the
  `widths` and `col.names` to maintain compatibility. #381
* Add (back) a pkgdown website:
  [https://gesistsa.github.io/rio/](https://gesistsa.github.io/rio/)
* Update all test cases #380
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants