1- v3.11.13 (XXXX-XX-XX )
1+ v3.11.13 (2025-02-07 )
22---------------------
33
4- * Disable cluster overwhelm protection on agents. This fixes a bug which
5- happens if the number of dbservers and coordinators grows beyond 4x
6- the number of threads in the agency leader.
4+ * Disable cluster overwhelm protection on agents. This fixes a bug which happens
5+ if the number of dbservers and coordinators grows beyond 4x the number of
6+ threads in the agency leader.
77
8- * Improve geo index performance in the cluster with multiple shards. This
9- fixes BTS-2046. An unnecessary and bad SORT node is removed from the
10- query plan in the case that a geo index is used.
8+ * Improve geo index performance in the cluster with multiple shards. This fixes
9+ BTS-2046. An unnecessary and bad SORT node is removed from the query plan in
10+ the case that a geo index is used.
1111
1212* BTS-2074: fix installing Foxx service from zip file upload using the web UI
1313 with multi-coordinator load balancing.
@@ -23,34 +23,32 @@ v3.11.13 (XXXX-XX-XX)
2323* Rebuilt included rclone v1.62.2 with go1.22.10 and non-vulnerable
2424 dependencies.
2525
26- * Updated ArangoDB Starter to v0.18.12 and arangosync to
27- v2.19.12.
26+ * Updated ArangoDB Starter to v0.18.12 and arangosync to v2.19.12.
2827
29- * Introduce expiry time for idle TCP/IP connections in the ConnectionCache
30- (for `SimpleHttpClient`) with a default of 120s. This is to prevent
31- errors in replication caused by cloud environments terminating
32- connections. Also add retries in a few places. Also increase the
33- timeout in initial sync to transfer up to 5000 documents from 25s to
34- 900s. This addresses BTS-2011, BTS-2042 and BTS-2035.
28+ * Introduce expiry time for idle TCP/IP connections in the ConnectionCache (for
29+ `SimpleHttpClient`) with a default of 120s. This is to prevent errors in
30+ replication caused by cloud environments terminating connections. Also add
31+ retries in a few places. Also increase the timeout in initial sync to transfer
32+ up to 5000 documents from 25s to 900s. This addresses BTS-2011, BTS-2042 and
33+ BTS-2035.
3534
3635* Don't cleanup failed agency jobs if they are subjobs of pending jobs.
3736 This avoids a bug in CleanOutServer jobs, where such a job could complete
38- seemingly successfully despite the fact that some MoveShard jobs had
39- actually failed. This fixes BTS-2022.
37+ seemingly successfully despite the fact that some MoveShard jobs had actually
38+ failed. This fixes BTS-2022.
4039
4140* Upgraded OpenSSL to 3.4.0.
4241
43- * Fix shard synchronisation race where after a shard move the new
44- leader informs followers before it updates Current in the Agency. In
45- some cases the old leader fell subsequently out of sync.
42+ * Fix shard synchronisation race where after a shard move the new leader informs
43+ followers before it updates Current in the Agency. In some cases the old
44+ leader fell subsequently out of sync.
4645
4746* BTS-2017: fix CleanOutServer for satellite collections.
4847
4948* BTS-2014: Fix delay if write hits early after a hotbackup restore.
5049
51- * Fix leader resignation race in coordinator, which lead to forgotten
52- collection read locks on dbservers, which in turn could lead to deadlocks
53- in the cluster.
50+ * Fix leader resignation race in coordinator, which lead to forgotten collection
51+ read locks on dbservers, which in turn could lead to deadlocks in the cluster.
5452
5553
5654v3.11.12 (2024-10-31)
0 commit comments