Bug #40213
closed
mds: cannot switch mds state from standby-replay to active
Added by simon gao almost 5 years ago.
Updated over 4 years ago.
Description
if a standby-replay mds run for a long time, there are too many inodes in cache. In the rejoin phase, the mds server traverses all nodes.
so we will trim the cache of standby-replay mds when some segments has been trimmed.
- Project changed from rgw to CephFS
- Subject changed from cephfs: can't switch mds state from standby-replay to active to mds: cannot switch mds state from standby-replay to active
- Status changed from New to Fix Under Review
- Assignee set to simon gao
- Priority changed from Normal to Urgent
- Start date deleted (
06/10/2019)
- Affected Versions v12.2.12 added
- Affected Versions deleted (
v12.2.10)
- Component(FS) MDS added
- Status changed from Fix Under Review to Pending Backport
- Backport changed from nautilus,mimic,luminous to nautilus,mimic
- Copied to Backport #40440: nautilus: mds: cannot switch mds state from standby-replay to active added
- Copied to Backport #42375: mimic: mds: cannot switch mds state from standby-replay to active added
- Status changed from Pending Backport to Resolved
- Related to Bug #48673: High memory usage on standby replay MDS added
Also available in: Atom
PDF