1.19.0.post2 “Fantastic Donkeys”¶
Upgrade Instructions¶
This feature release requires a database schema upgrade. Please consult the documentation about upgrading your database schema.
The following changes are necessary and are covered by the alembic upgrade head
command:
bad_replicas_table
changes (Alembic revisionb96a1c7e1cc4
)New
expires_at
column.Primary key change.
New
BAD_REPLICAS_EXPIRES_AT_IDX
Index.Constraint
BAD_REPLICAS_STATE_CHK
change.
New
bad_pfns
table (Alembic revisionb96a1c7e1cc4
)Change of
rse_attr_map
andaccount_attr_map
boolean values (Alembic revision9eb936a81eb1
)True
totrue
False
tofalse
Changes in history tables (Alembic revision
bf3baa1c1474
)Drop of Primary Key
CONTENTS_HIST_PK
for tablecontents_history
.Drop of Primary Key
ARCH_CONTENTS_HISOTRY_PK
for tablearchive_contents_history
.Creation of Index
ARCH_CONT_HIST_IDX
for tablearchive_contents_history
.Drop of Primary Key
RULES_HIST_RECENT_PK
for tablerules_hist_recent
.Drop of column
history_id
for tablerules_hist_recent
.Drop of Primary Key
RULES_HIST_LONGTERM_PK
for tablerules_history
.Drop of column
history_id
for tablerules_history
.Drop of Primary Key
MESSAGES_HIST_ID_PK
for tablemessages_history
.
Adding of new
payload_nolimit
column to messages tables (Alembic revision90f47792bb76
)Add column
payload_nolimit
to tablemessages
.Add column
payload_nolimit
to tablemessages_history
.
Change of constraint
RULES_NOTIFICATION_CHK
forrules
(Alembic revision01eaf73ab656
)Change of Primary Key
REPLICAS_PK
for tablereplicas
(Alembic revision3345511706b8
)
The following change are necessary and are NOT covered by the alembic upgrade head
command:
The boolean values upgrade on the
rse_attr_map
andaccount_attr_map
tables from0
tofalse
and1
totrue
need to be executed manually, since the script cannot distinguish between the integer or boolean interpretation of the attribute. It is very important that you only upgrade0
and1
which actualy correspond to boolean values, otherwise functionality might be degraded (Alembic revision9eb936a81eb1
)
General¶
Features¶
Core & Internals: Mark replicas as temporarily unavailable #1550
Core & Internals: PK in history tables mostly wrong #1818
Core & Internals: Limit the maximum space usage on an RSE #1965
Core & Internals: Messages payload are limited to 4000 characters and throw exception #48
Core & Internals: json based import for RSEs, protocols, distances #716
Rules: New Rule Notification Mode #1922
Transfers: Remove conveyor-transfer-submitter #1632
Enhancements¶
Clients: Allow upload of replicas to non-det RSEs by users #1172
Core & Internals: Core migration to python3 #1912
Core & Internals: Daemon migration to python3 #1924
Core & Internals: fix last python3 incompatibilities #1961
Core & Internals: Add missing wsgi alias for import/export #1990
Messaging: improve connection handling to broker #1479
Rebalancing: bb8 - replacing sql query with sql alchemy #1861
Rebalancing: BB8 should avoid reducing the replication factor of a dataset #484
Release management: Address security issue in paramiko due to CVE-2018-1000805 #1646
Release management: Upgrade of dependencies for 1.19.0 #2028
Release management: Rucio py3 compatibility #67
Testing: switch oracle image for travis tests #2010
Traces: kronos: activemq connect is missing wait=True #321
Bugs¶
Core & Internals: Primary Key for replicas table is in wrong order in models.py #1749
Core & Internals: list-rse-usage with –show-accounts does not work #2040
Deletion: The reaper may ignore some rses if there is more than one rse handled per thread #1729
Deletion: Reaper crashes if no RSEs are defined #1917
Documentation: Documentation build fails #1774
Documentation: Broken links in documentation sidebar #2065
Probes & Alarms: The FTS probe sometimes doesn’t whitelist a SE after the end of downtime #2015
Probes & Alarms: Synchronisation with VOMS fails under Nagios due to non-ASCII characters #2051
Rebalancing: BB8 crash with RSEBlacklisted exception #313
Release management: Define stricter depdency ranges for pip-requires-client #1692
Transfers: REST call for requests is broken by ‘/’ in DIDs #1791