Highlights
-
allows the server to start under situations where the disk is faulty, this is to ensure that other
good drive(s) locally on the same node are usable. Previously we kept the servers down
simply to provide more attention for the administrator to fix these drives, but in containerized
environments this is not true and pods crashing can have other issues - to improve the
situation for containerized environments allow the server to be operable - instead the administrator
should look atmc admin infoto look for faulty drives to be fixed/replaced, see #10175 -
CopyObject optimization to avoid creating a new copy of the entire object if the destination
is the same object copied from the same source object with versionId specified. #10170 -
Allow write quorum calculation under reduced parity uniform across all write operations #10186
Changelog
- fix: proxies set keep-alive timeouts to be system dependent (#10199) (08/04/20) (Harshavardhana)
- fix: reduce an extra system call for writes instead fail later (#10187) (08/04/20) (Harshavardhana)
- fix: multi-tenant deployment guide to use ellipses format (#10192) (08/04/20) (Harshavardhana)
- fix ILM should not remove a protected version (#10189) (08/04/20) (Anis Elleuch)
- allow server to start even with corrupted/faulty disks (#10175) (08/03/20) (Harshavardhana)
- add CopyObject optimization when source and destination are same (#10170) (08/03/20) (Harshavardhana)
- fix: allow FS mode situations when conflicting files exist (#10185) (08/03/20) (Harshavardhana)
- fix: quorum calculation mistake with reduced parity (#10186) (08/03/20) (Harshavardhana)
- fix: connection_string should override other params (#10180) (08/03/20) (Harshavardhana)
- fix: pull credits in docker from master branch (#10177) (08/02/20) (Harshavardhana)
- fix: update and improve zh_CN documentation (#10174) (08/02/20) (飞雪无情)
- update CREDITS with minio/selfupdate project (07/31/20) (Harshavardhana)