The previous cap was set at 1 MB. However, a user was experiencing mon
timeouts while syncing the purged_snap_epoch * keys, which are ~20 bytes
each. Reducing the max payload to 64K resolved the problem, which maps
to (very!) roughly 1500 keys per message. Set our limit a bit higher than
that since we just made this quite a bit more efficient. Most of the time
the keys are larger than 20 bytes and we wouldn't hit the key limit, but
having one ensures that we won't burn too much CPU in one go when we do
have lots of these little keys.
Signed-off-by: Sage Weil <sage@redhat.com>
This commit updates the "Install NTP" section of the preflight
checklist, instructing readers who are using Ubuntu 18.04 to
use the package name "chrony" where elsewhere the package name
"ntp" would be used. "chrony" has replaced "ntp" in Ubnutu 18.04.
This commit fixes Bug number 8 in the list here:
https://pad.ceph.com/p/Report_Documentation_Bugs
Signed-off-by: Zac Dover <zac.dover@gmail.com>
mgr/dashboard: fix tests in order to match pg num conventions
Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
This will ensure that the Ceph dashbord's block mirroring page and
the CLI's 'mirror pool status' have matching health indications.
Fixes: https://tracker.ceph.com/issues/42748
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
* refs/pull/31797/head:
qa: test mgr cephfs mount blacklist
mgr: forward RADOS client instances for potential blacklist
mon/MgrMonitor: blacklist previous instance
mon: set RUNTIME flag on mon_mds_blacklist_interval
mgr: use more efficient data structure
Reviewed-by: Sage Weil <sage@redhat.com>
by adding the previously added monitoring related features as well as
the newest feature addition. Extends the documentation where necessary
to describe the Prometheus' alert configuration.
Fixes: https://tracker.ceph.com/issues/42877
Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
tools/osdmaptool.cc: do not use deprecated std::random_shuffle()
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>
The mgr creates a per-module RADOS client connection for modules which
interact with RADOS (e.g. the volumes module). These clients should also
be blacklisted when the active mgr is failed; we don't want the former
active mgr to continue interacting with RADOS when the new one takes
over. This is particularly impactful for avoiding extraneous
"unresponsive client" warnings from the MDS when the mgr switches
(especially in testing). The MDS will pickup the new OSD blacklists
which include's the old mgr's libcephfs instance and blacklist/evict
that session quietly.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
This wasn't realized as necessary early on in the ceph-mgr development
because the mgr didn't interact with RADOS. However, now it is becoming
common for plugins to store data there. It's important that the previous
instance can no longer interact with RADOS while the new mgr takes over.
In particular, this means that the mgr's client sessions with the MDS
are automatically evicted once the MDS receives the new OSDMap. This
avoids a pesky "unresponsive client" warning in the cluster logs.
Fixes: https://tracker.ceph.com/issues/42939
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
the use of `std::random_shuffle()` was introduced by
b946308 .
in this change, it is replaced using `std::shuffle()`.
Fixes: https://tracker.ceph.com/issues/43084
Signed-off-by: Kefu Chai <kchai@redhat.com>
If we start the command before we get the mgrmap at the client, we
will send MCommand via a different code path; make sure it behaves
correctly too.
Missed this in 6f35d28352
Signed-off-by: Sage Weil <sage@redhat.com>
It is normal to set things in teh mon config database that can't be absorbed
by clients/daemons at runtime. Things like ms_type can/should be set there
(so that they, for instance, help produce the 'config minimal-conf') but
we don't need daemons to spit out error messages every time they get a config
update.
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/31869/head:
ceph-daemon: bootstrap: deploy initial mon via deploy_daemon()
qa/standalone/test_ceph_daemon.sh: more $SUDO
ceph-daemon: configure firewalld for new daemon deploys
ceph-daemon: name mgr the same way mgr/ssh does
Reviewed-by: Michael Fritch <mfritch@suse.com>
mgr/dashboard: grafana charts match time picker selection.
Reviewed-by: Jan Fajerski <jfajerski@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Patrick Seidensal <pnawracay@suse.com>
Note that we only open firewalld ports; we never close them.
We could conceivably do that from rm-cluster, but that might also interfere
with something else on the host...
Signed-off-by: Sage Weil <sage@redhat.com>
with a tab for 'active alerts', 'all alerts' and 'silences'. Due to
ambiguity with existing names, `AlertListComponent` has been renamed to
`ActiveAlertListComponent`. Introduces `MonitoringListComponent` as
first page for monitoring concerns, using path `/monitoring`.
Keeps the activated tab open, independent of the way that's used to go
back to the previous page, be it the cancel button or submit button or
the link on the breadcrumb. Also keeps the active tab open even when the
page is reloaded.
Fixes: https://tracker.ceph.com/issues/42877
Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
Mocking the test the way it was removed the asynchronous nature of the
test. By using an Observable the test can stay asynchronous and be
tested as well.
Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
