profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/batrick/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Patrick Donnelly batrick Red Hat, Inc. New Hampshire, USA http://www.batbytes.com/ CephFS Team Lead

batrick/lunadry 28

A code formatter in Lua.

batrick/lua-signal 8

A signal library for Lua.

batrick/ceph-linode 6

Launch Ceph using the Linode VPS provider

batrick/lua-serialize 5

Pure Lua serialization library.

batrick/lua-structure 1

A module for checking that a table conforms to a strictly typed structure.

batrick/basekernel 0

A simple OS kernel for research, teaching, and fun.

batrick/cctools 0

The Cooperative Computing Tools (cctools) enable large scale distributed computations to harness hundreds to thousands of machines from clusters, clouds, and grids.

batrick/ceph 0

Ceph is a distributed object, block, and file storage platform

batrick/ceph-ansible 0

Ansible playbooks for Ceph

pull request commentceph/ceph

pybind/mgr/cephadm: set allow_standby_replay during CephFS upgrade

I think the more serious issue is the current upgrade logic only performs scale down when upgrading across major versions. So this time when upgrading from 16.2.5 to 16.2.6, the down-scale logic is not triggered, and all new MDS cannot be promoted from standby.

Thanks for pointing that out. That will need fixed too!

batrick

comment created time in 18 hours

push eventbatrick/ceph

Patrick Donnelly

commit sha 7344629ed0e324f73ea978d0324c73bc828f739d

pybind/mgr/cephadm: set allow_standby_replay during CephFS upgrade Following procedure in [1]. Also: harden checks for active. Ensure "up" and "in" are both [0]. There should be no standby-replay daemon. [1] https://docs.ceph.com/en/pacific/cephfs/upgrading/ Fixes: https://tracker.ceph.com/issues/52654 Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

push time in a day

push eventbatrick/ceph

Patrick Donnelly

commit sha 051ddc3dff86d3a13aa2109015e0a2bc1eef1822

pybind/mgr/cephadm: set allow_standby_replay during CephFS upgrade Following procedure in [1]. Also: harden checks for active. Ensure "up" and "in" are both [0]. There should be no standby-replay daemon. [1] https://docs.ceph.com/en/pacific/cephfs/upgrading/ Fixes: https://tracker.ceph.com/issues/52654 Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

push time in a day

PR opened ceph/ceph

Reviewers
pybind/mgr/cephadm: set allow_standby_replay during CephFS upgrade cephfs cephadm

This will need some kind of QA test. Would like some direction on that part.

<!-- Thank you for opening a pull request! Here are some tips on creating a well formatted contribution.

Please give your pull request a title like "[component]: [short description]"

This is the format for commit messages:

""" [component]: [short description]

[A longer multiline description]

Fixes: [ticket URL on tracker.ceph.com, create one if necessary] Signed-off-by: [Your Name] <[your email]> """

The Signed-off-by line is important, and it is your certification that your contributions satisfy the Developers Certificate or Origin. For more detail, see SubmittingPatches.rst.

The component is the short name of a major daemon or subsystem, something like "mon", "osd", "mds", "rbd, "rgw", etc. For ceph-mgr modules, give the component as "mgr/<module name>" rather than a path into pybind.

For more examples, simply use "git log" and look at some historical commits.

This was just a quick overview. More information for contributors is available here: https://raw.githubusercontent.com/ceph/ceph/master/SubmittingPatches.rst

-->

Checklist

  • [ ] References tracker ticket
  • [ ] Updates documentation if necessary
  • [ ] Includes tests for new functionality or reproducer for bug

<details> <summary>Show available Jenkins commands</summary>

  • jenkins retest this please
  • jenkins test classic perf
  • jenkins test crimson perf
  • jenkins test signed
  • jenkins test make check
  • jenkins test make check arm64
  • jenkins test submodules
  • jenkins test dashboard
  • jenkins test dashboard cephadm
  • jenkins test api
  • jenkins test docs
  • jenkins render docs
  • jenkins test ceph-volume all
  • jenkins test ceph-volume tox

</details>

+75 -32

0 comment

1 changed file

pr created time in a day

create barnchbatrick/ceph

branch : i52654

created branch time in a day

create barnchceph/ceph-ci

branch : ceph-mds-addfailed-pacific

created branch time in 2 days

create barnchceph/ceph-ci

branch : wip-pdonnell-ceph-mds-addfailed

created branch time in 2 days

create barnchceph/ceph-ci

branch : wip-pdonnell-testing-20210917.174826

created branch time in 2 days

push eventceph/ceph

wangxinyu

commit sha f0d1635f26f74d21ddb9be92570547dce924efdf

doc/cephfs: fix typos Signed-off-by: wangxinyu <wangxinyu@inspur.com>

view details

Patrick Donnelly

commit sha e81a0c72a4b9c9c2ae97b22223b9f2bfd101fc35

Merge PR #43144 into master * refs/pull/43144/head: doc/cephfs: fix typos Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

view details

push time in 2 days

PR merged ceph/ceph

doc/cephfs: fix typos documentation cephfs

Signed-off-by: wangxinyu wangxinyu@inspur.com

<!-- Thank you for opening a pull request! Here are some tips on creating a well formatted contribution.

Please give your pull request a title like "[component]: [short description]"

This is the format for commit messages:

""" [component]: [short description]

[A longer multiline description]

Fixes: [ticket URL on tracker.ceph.com, create one if necessary] Signed-off-by: [Your Name] <[your email]> """

The Signed-off-by line is important, and it is your certification that your contributions satisfy the Developers Certificate or Origin. For more detail, see SubmittingPatches.rst.

The component is the short name of a major daemon or subsystem, something like "mon", "osd", "mds", "rbd, "rgw", etc. For ceph-mgr modules, give the component as "mgr/<module name>" rather than a path into pybind.

For more examples, simply use "git log" and look at some historical commits.

This was just a quick overview. More information for contributors is available here: https://raw.githubusercontent.com/ceph/ceph/master/SubmittingPatches.rst

-->

Checklist

  • [ ] References tracker ticket
  • [ ] Updates documentation if necessary
  • [ ] Includes tests for new functionality or reproducer for bug

<details> <summary>Show available Jenkins commands</summary>

  • jenkins retest this please
  • jenkins test classic perf
  • jenkins test crimson perf
  • jenkins test signed
  • jenkins test make check
  • jenkins test make check arm64
  • jenkins test submodules
  • jenkins test dashboard
  • jenkins test dashboard cephadm
  • jenkins test api
  • jenkins test docs
  • jenkins render docs
  • jenkins test ceph-volume all
  • jenkins test ceph-volume tox

</details>

+2 -2

0 comment

2 changed files

wxypro

pr closed time in 2 days

PullRequestReviewEvent

pull request commentceph/ceph

mds: do not take the ino which has been used

Xiubo is looking into adding a test for this.

lxbsz

comment created time in 2 days

issue commentrook/rook

filesystem is offline on 16.2.6 deployment

Upgrading requires you to now disable allow_standby_replay: https://docs.ceph.com/en/pacific/cephfs/upgrading/#upgrading-the-mds-cluster

NAB. This is by design.

leseb

comment created time in 2 days

pull request commentceph/ceph

qa: convert fs:workload to use cephadm

What do you think of having a super small subset of this test here in the orch/cephadm suite? Maybe two jobs or so and then let the cephadm team keep this subset stable?

It's okay with me.I'll try to get around to it when I have some time.

batrick

comment created time in 2 days

PR opened ceph/ceph

pacific: MDSMonitor: handle damaged state from standby-replay cephfs needs-qa pacific-batch-1

https://tracker.ceph.com/issues/52639

+123 -65

0 comment

11 changed files

pr created time in 2 days

create barnchbatrick/ceph

branch : i52639

created branch time in 2 days

push eventceph/ceph

Patrick Donnelly

commit sha 976741a8f5f0e7a7d0b10fd75f0cce99a73a221c

mds: harden rank lookup This is a trivial refactor. Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 4034fc7c6acbbba90d07e52f4bcd8b6738e47b8c

mds: refactor iterator lookup Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 9785a3623190b6491f88181ba6b6a876d3949d5e

include: unset std::hex after printing CompatSet Otherwise you see this in debug output: $ ceph fs dump ... standby_count_wanted 0 [mds.d{0:4375} state up:standby-replay seq 1 addr [v2:127.0.0.1:6830/2171112379,v1:127.0.0.1:6831/2171112379] compat {c=[1],r=[1],i=[7ff]}] Standby daemons: [mds.b{ffffffff:110a} state up:standby seq 1 addr [v2:127.0.0.1:1aab/d4c5b653,v1:127.0.0.1:1aad/d4c5b653] compat {c=[1],r=[1],i=[7ff]}] Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 391f3c8790e58d7f67b7bf93665775d5c4191b8c

mds: add config to mark rank damaged in standby-replay For testing. Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 20509bb6c82e872127ab838d45402be0d0b91b5f

MDSMonitor: handle damaged from standby-replay This change does a few things: - if a state transition is invalid or a beacon is garbage, the MDSMonitor now evicts the MDS instead of ignoring the problem. - standby state validation is moved to prepare_beacon where eviction can happen. - standby-replay may indicate the rank is damaged (failure to replay the journal). - if the rank is damaged, both the rank holder and standby-replay daemon (if any) will be removed. Fixes: https://tracker.ceph.com/issues/52565 Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha f4a11a329020d31a3f99a5099201676d3e1fb2b9

qa: add test for standby-replay marking rank damaged Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 0d9753fa3c1a7b5ae35989b9f6cda17af79cffbb

Merge PR #43122 into master * refs/pull/43122/head: qa: add test for standby-replay marking rank damaged MDSMonitor: handle damaged from standby-replay mds: add config to mark rank damaged in standby-replay include: unset std::hex after printing CompatSet mds: refactor iterator lookup mds: harden rank lookup Reviewed-by: Venky Shankar <vshankar@redhat.com>

view details

push time in 2 days

delete branch batrick/ceph

delete branch : i52565

delete time in 2 days

PR merged ceph/ceph

MDSMonitor: handle damaged from standby-replay bug fix documentation core cephfs tests common needs-review mon wip-pdonnell-testing
+124 -65

1 comment

11 changed files

batrick

pr closed time in 2 days

Pull request review commentceph/ceph

qa: fsync dir for asynchronous creat on stray tests

 def open_n_background(self, fs_path, count):         self.background_procs.append(rproc)         return rproc -    def create_n_files(self, fs_path, count, sync=False):+    def create_n_files(self, fs_path, count, sync=False, dirsync=False, unlink=False, finaldirsync=False):

fixed, thanks!

batrick

comment created time in 2 days

PullRequestReviewEvent

push eventbatrick/ceph

Patrick Donnelly

commit sha 8a4eef85c0d99aa0bf7c74d6d325e954f8100f75

qa: refactor and generalize create_n_files Few things: - Allow calling fsync on directory (to support async create kernel). - Allow immediately unlinking the created file (for stray testing). - Close any file descriptors created. - Write unique content (the i variable) to each file. Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha c6892287b0889f192b7824fde9b500ba805c53cc

qa: fsync dir for asynchronous creat on stray tests Use the enhanced create_n_files to dedup code. Also split the large test into three. Fixes: https://tracker.ceph.com/issues/52606 Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

push time in 2 days

Pull request review commentceph/ceph

qa: fsync dir for asynchronous creat on stray tests

 def open_n_background(self, fs_path, count):         self.background_procs.append(rproc)         return rproc -    def create_n_files(self, fs_path, count, sync=False):+    def create_n_files(self, fs_path, count, sync=False, dirsync=False, unlink=False, finaldirsync=False):

done

batrick

comment created time in 2 days

PullRequestReviewEvent

push eventbatrick/ceph

Venky Shankar

commit sha 746278727e10c9a2b28f0f9a43401ed88d23de70

mgr/mirroring: `daemon status` command does not require file system name Signed-off-by: Venky Shankar <vshankar@redhat.com>

view details

Venky Shankar

commit sha 11b61b4fb91f731ee1d04cc545bfa4937e844bb5

test: adjust mirroring tests for `daemon status` change Signed-off-by: Venky Shankar <vshankar@redhat.com>

view details

Venky Shankar

commit sha ed253597c9376ac76171ca27bece09be8a72e269

doc: fix `daemon status` interface (exclude file system name) Fixes: http://tracker.ceph.com/issues/51989 Signed-off-by: Venky Shankar <vshankar@redhat.com>

view details

Xiubo Li

commit sha cf9d48ab7f32f9a4f17366cf3abf61d18e3984cd

mds: switch to use ceph_assert() instead of assert() If the -DNDEBUG was specified when build the code, the assert() will do nothing. Signed-off-by: Xiubo Li <xiubli@redhat.com>

view details

Xiubo Li

commit sha 56e92ee434d75e9be82217885ef5030f7bb3768c

client: switch to use ceph_assert instead of assert If the -DNDEBUG was specified when build the code, the assert() will do nothing. Signed-off-by: Xiubo Li <xiubli@redhat.com>

view details

Jeff Layton

commit sha 9c33a590466f7c553ea6bc62e90dc67728fa1e26

mds: ensure that we send the btime in cap messages While troubleshooting a different problem, I noticed that the kclient would occasionally not report the btime in statx calls. It's most evident in the case where two different clients are alternately writing to the same file, and another task on one of them is repeatedly issuing statx() calls against it. The problem is that we're not correctly populating the MClientCaps btime field when issuing caps, and that causes the MDS to incorrectly send a zeroed-out btime field. Fixes: https://tracker.ceph.com/issues/52123 Signed-off-by: Jeff Layton <jlayton@redhat.com>

view details

Kotresh HR

commit sha e5b6737db1f57dda8c6f1229286136e574633765

mds: Add new flag to MClientSession The "error_string" in the metadata of MClientSession is being parsed by kclient to validate whether the session is blocklisted. The "error_string" is for humans and shouldn't be relied on it. Hence added the flag to MClientsession to indicate the session is blocklisted. Signed-off-by: Kotresh HR <khiremat@redhat.com> Fixes: https://tracker.ceph.com/issues/52382

view details

Aashish Sharma

commit sha 58d635455d1f59921d5ad821168f31b6f937588a

mgr/dashboard: Incorrect MTU mismatch warning The MTU mismatch warning was being fired for those NIC's as well that are in down state. This PR intends to fix this issue Fixes:https://tracker.ceph.com/issues/52028 Signed-off-by: Aashish Sharma <aasharma@redhat.com>

view details

Patrick Donnelly

commit sha 1a9aab62b615830107ee28866630f043e1a01a35

mon: refactor quorum age calculation Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Patrick Donnelly

commit sha 89c901a4944158c21eb26d53676a709fa2964c9d

mon: do not quickly mark mds laggy when MON_DOWN The MDS may be sending beacons to a partitioned or newly restarted monitor. This will wrongly cause the current leader to believe the MDS is unavailable. Fixes: https://tracker.ceph.com/issues/43216 Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

view details

Kotresh HR

commit sha 7440ef842a41eff3ac4751c487dacc5df2603095

mgr/volumes: Fix permission during subvol creation with mode The subvolume creation with specific mode leads to creation of parent directories ('/volumes/_no_group') with the same mode if it's not already created. Fixed the same. Similarly, the subvolumegroup creation with specific mode leads to creation of parent directory ('/volumes') with same mode if it's not already created. Fixed the same. Fixes: https://tracker.ceph.com/issues/51870 Signed-off-by: Kotresh HR <khiremat@redhat.com>

view details

Radoslaw Zarzynski

commit sha 093586ff313db9c12e5b4d1b9eb5210a8adf8aa8

crimson/net: simplify error handling in FixedCPUServerSocket::listen(). Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

view details

Radoslaw Zarzynski

commit sha 3b63f4c85c682db13b309233063c057fd01b1289

crimson/net: s/do_bind/do_listen/ in SocketMessenger. Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

view details

Yongseok Oh

commit sha 49270bc6d2fcab2daaa729ca2b14e692152c82ff

mds: add mds_dir_max_entries config option fixes: https://tracker.ceph.com/issues/52491 Signed-off-by: Yongseok Oh <yongseok.oh@linecorp.com>

view details

Yongseok Oh

commit sha fed9aba38caa779ed571345dd6afc5f81cedb2b7

doc/cephfs: add mds_dir_max_entries option into mds-config-ref.rst Signed-off-by: Yongseok Oh <yongseok.oh@linecorp.com>

view details

Michael Fritch

commit sha fd977773a57e12003fb02bdc762bf6bc89d785a1

cephadm: skip podman check during `rm-repo` allow the `rm-repo` command to succeed when podman is not installed Signed-off-by: Michael Fritch <mfritch@suse.com>

view details

Jianwei Zhang

commit sha dcdb188b6f577551fb377ba34145419f81322b03

osd: fix partial recovery become whole object recovery after restart osd support SERVER_OCTOPUS feature for pg_missing_item::encode() Fixes: https://tracker.ceph.com/issues/52583 Signed-off-by: Jianwei Zhang <jianwei1216@qq.com>

view details

Joseph Sawaya

commit sha 58d13a5dc44731b3f33450896790dd33f4c2bd53

mgr/rook: translate placement spec to node selector and vice versa This commit creates methods for translating PlacementSpecs to NodeSelectors and vice versa, this is a general method that would be used by orch commands that have a placement option. It's able to translate specs that specify a label, hosts and the host_pattern '*'. Signed-off-by: Joseph Sawaya <jsawaya@redhat.com>

view details

Joseph Sawaya

commit sha 5d734210797663e86f9f91aeaab5f28184bcd01c

mgr/rook: add unit tests to Rook module This commit creates a unit test folder for rook, configurates tox and creates a unit test to test the translation of placement specs to node selectors and vice versa. Signed-off-by: Joseph Sawaya <jsawaya@redhat.com>

view details

Radoslaw Zarzynski

commit sha 7f85b51161ababa843dc64beb880b53bc02ffc85

crimson/net: decouple listen_ertr from FixedCPUServerSocket. Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

view details

push time in 2 days

Pull request review commentceph/ceph

qa: fsync dir for asynchronous creat on stray tests

 def open_n_background(self, fs_path, count):         self.background_procs.append(rproc)         return rproc -    def create_n_files(self, fs_path, count, sync=False):+    def create_n_files(self, fs_path, count, sync=False, dirsync=False, unlink=False, finaldirsync=False):         assert(self.is_mounted())          abs_path = os.path.join(self.hostfs_mntpt, fs_path) -        pyscript = dedent("""-            import sys-            import time+        pyscript = dedent(f"""             import os+            from contextlib import closing

nowhere, I was going to use it but found out it doesn't work.

batrick

comment created time in 2 days

PullRequestReviewEvent
PullRequestReviewEvent
CommitCommentEvent
CommitCommentEvent