pg_autoscale module will now start out all the pools
with a scale-up profile by default.
Added tests in workunits/mon/pg_autoscaler.sh
to evaluate if the default pool creation is
a scale-up profile
Updated documentation and release notes to
reflect the change in the default behavior
of the pg_autoscale profile.
Fixes: https://tracker.ceph.com/issues/53309
Signed-off-by: Kamoltat <ksirivad@redhat.com>
the loop of proc.communicate() on python3.6, where we always are able to
get something out of stdout and/or stderr PIPEs. and the `stdout` and
`stderr` keep growing until out of memory. and teuthology considers
the command crashed after a while.
Fixes: https://tracker.ceph.com/issues/50393
Signed-off-by: Kefu Chai <kchai@redhat.com>
The autoscaler will start out with scaling each
pools to have a full complements of pgs from the start
and will only decrease it when pools need more due to
increased usage.
Introduced a unit test that tests only the
function get_final_pg_target_and_ratio() which
deals with the distrubtion of pgs amongst the
pools
Edited workunit script to reflect the change
of how pgs are calculated and distrubted.
Signed-off-by: Kamoltat <ksirivad@redhat.com>
verify whether min_size is recalculated when osd
pool size is changed.
fixes: https://tracker.ceph.com/issues/44862
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
Adds option `mon_allow_pool_size_one` which will be disabled by default
to ensure pools are not configured without replicas.
If the user still wants to use pool size 1, they will have to change the
value of `mon_allow_pool_size_one` to true and then have to pass flag
`--yes-i-really-mean-it` to cli command:
Example:
`ceph osd pool test set size 1 --yes-i-really-mean-it`
Fixes: https://tracker.ceph.com/issues/44025
Signed-off-by: Deepika Upadhyay <dupadhya@redhat.com>
Also check for pg_num_target being set correctly, rather than pg_num,
so the test doesn't depend on merging/splitting speed.
Signed-off-by: Josh Durgin <jdurgin@redhat.com>
Normalize across pools so that it's simpler to use - this way you
don't have to adjust every other pool when you add one.
Handle pools with target_bytes by taking their capacity off the top,
and dividing the rest into the pools with a target_ratio.
If both target bytes and ratio are specified, ignore bytes. This
matches the docs and makes accounting simpler.
Fixes: https://tracker.ceph.com/issues/43947
Signed-off-by: Josh Durgin <jdurgin@redhat.com>
These options are special. Since we can't get the values from the mon,
we also shouldn't store them there. Most of them are not options that
can be changed anyway.
Signed-off-by: Sage Weil <sage@redhat.com>
To be able to catch problems with python2 *and* python3, run flake8
with both versions. From the flake8 homepage:
It is very important to install Flake8 on the correct version of
Python for your needs. If you want Flake8 to properly parse new
language features in Python 3.5 (for example), you need it to be
installed on 3.5 for Flake8 to understand those features. In many
ways, Flake8 is tied to the version of Python on which it runs.
Also fix the problems with python3 on the way.
Note: This requires now the six module for teuthology. But this is
already an install_require in teuthology itself.
Signed-off-by: Thomas Bechtold <tbechtold@suse.com>
There were a couple of problems found by flake8 in the qa/
directory (most of them fixed now). Enabling flake8 during the usual
check runs hopefully avoids adding new issues in the future.
Signed-off-by: Thomas Bechtold <tbechtold@suse.com>
For an entity a.b.c.d, search all dot-delineated prefix sections. This
enables you to establish a hierarchical set of options for clients, such
as radosgw daemons.
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/30859/head:
auth: EACCES, not EPERM
mon: shunt old tell commands from cli interface to asok
mon: allow mgr to tell mon.foo smart
mon: include quorum features in quorum_status
qa/workunits/mon/caps.sh: fix test
ceph_test_rados_api_cmd: fix MonDescribe test
Merge branch 'vstart-fs-auth' of git://github.com/batrick/ceph into wip-cleanup-mon-asok
test/pybind/test_ceph_argparse: fix tests
vstart: add volume client keys to keyring
vstart: use fs authorize to create master client key
vstart: redirect some output to stderr
vstart: output command strings to stderr
qa/workunits/cephtool/test.sh: fix 'quorum enter' caller
qa: change mon_status calls to quorum_status or tell commands
mon: fix 'heap ...' command
mon: consolidate 'sync force' commands
mon: allow asok commands to return an error code
mon: move 'quorum enter|exit' and 'mon_status' to asok
mon: fix 'smart' asok command
mon: remove old 'config set' and 'injectargs'
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
* refs/pull/30813/head:
qa: get rid of iteritems for python3 compatibility
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
I'm not really sure why this test expected EPERM before when it expects 0
a bit earlier, but it should certainly expect EPERM after the user is
deleted.
Signed-off-by: Sage Weil <sage@redhat.com>
This way it can be used to fast cancel/undo the last command.
Also make the tip message a litter bit nicer..
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
There is a lot of good stuff going on here, but nobody is investing in xio
and it is not expected to be the path forward for RDMA. If that ever
changes, we can resurrect the code. Until then, let's clean up the tree
and reduce friction for changes going forward.
Signed-off-by: Sage Weil <sage@redhat.com>
The crush/builder.c crush_add_bucket method resizes the max_buckets array
but a power of 2 when it has to expand, but the code in CrushWrapper was
assuming that if the array grew the pos for the new bucket would be the
last position in the new array. This led to a situation where the
crush_choose_arg_map args array size didn't match max_buckets, and
eventually caused a crash.
Fixes: http://tracker.ceph.com/issues/38664
Signed-off-by: Sage Weil <sage@redhat.com>
This avoids a huge pg merge from 100s to 4, which takes a long time and
makes the teuthology scrub cleanup time out.
Signed-off-by: Sage Weil <sage@redhat.com>
Henceforth, we'll require explicit `allow` caps for commands, or for the
config-key service. Blanket caps are no longer allowed for the
config-key service, except for 'allow *'.
Signed-off-by: Joao Eduardo Luis <joao@suse.de>
in b38b8e980c, we changed the upper
limit of size of `config key` 's value to 64k, so we need to update
the test accordingly.
Fixes: http://tracker.ceph.com/issues/36260
Signed-off-by: Kefu Chai <kchai@redhat.com>
5s -> 5m to gives us more leeway for when the mons are thrashing.
Also, *only* set this timeout when we expect a timeout. If we don't,
wait forever.
Signed-off-by: Sage Weil <sage@redhat.com>
