v1.3.5

Fixes

• Fix the issue that old nighlty may cause error (#1198, @lucklove)

v1.3.4

Fixes

• Fix the issue that tiup-cluster can't gernerate prometheus config (#1185, @lucklove)
• Fix the issue that tiup may choose yanked version if it's already installed (#1191, @lucklove)

v1.3.3

Fixes

• Fix the issue that tiup will hang forever when reloading a stopped cluster (#1044, @9547)
• Fix the issue that tiup mirror merge does not work on official offline package (#1121, @lucklove)
• Fix the issue that there may be no retry when download component failed (#1137, @lucklove)
• Fix the issue that PD dashboard does not report grafana address in playground (#1142, @9547)
• Fix the issue that the default selected version may be a preprelease version (#1128, @lucklove)
• Fix the issue that the error message is confusing when the patched tar is not correct (#1175, @lucklove)

Improvements

• Add darwin-arm64 not support hint in install script (#1123, @terasum)
• Improve playground welcome information for connecting TiDB (#1133, @dveeden)
• Bind latest stable grafana and prometheus in DM deploying (#1129, @lucklove)
• Use the advertised host instead of 0.0.0.0 for tiup-playground (#1152, @9547)
• Check tarball checksum on tiup-server when publish component (#1163, @lucklove)

v1.3.2

Fixes

• Fix the issue that the grafana and alertmanager target not set in prometheus.yaml (#1041, @9547)
• Fix the issue that grafana deployed by tiup-dm missing home.json (#1056, @lucklove)
• Fix the issue that the expires of cloned mirror is shourened after publish component to it (#1051, @lucklove)
• Fix the issue that tiup-cluster may remove wrong paths for imported cluster on scale-in (#1068, @AstroProfundis)
  • Risk of this issue: If an imported cluster has deploy dir ending with /, and sub dirs as <deploy-dir>//sub, it could results to delete wrong paths on scale-in
• Fix the issue that imported *_exporter has wrong binary path (#1101, @AstroProfundis)

Improvements

• Apply more strict check on tar.gz file for patch command: check if the entry is an executable file (#1091, @lucklove)

v1.3.1

Fixes

• Workaround the issue that store IDs in PDs may not monotonically assigned (#1011, @AstroProfundis)
  • Currently, the ID allocator is guaranteed not to allocate duplicated IDs, but when PD leader changes multiple times, the IDs may not be monotonic
    • For tiup < v1.2.1, the command tiup cluster display may delete store (without confirm) by mistake due to this issue (high risk)
    • For tiup >= v1.2.1 and <= v1.3.0, the command tiup cluster display may display up stores as tombstone, and encourages the user to delete them with the command tiup cluster prune (medium risk)
• Fix the issue that the cluster check always fail on thp check even though the thp is disabled (#1005, @lucklove)
• Fix the issue that the command tiup mirror merge -h outputs wrong usage (#1008, @lucklove)
  • The syntax of this command should be tiup mirror merge <mirror-dir-1> [mirror-dir-N] but it outputs tiup mirror merge <base> <mirror-dir-1> [mirror-dir-N]
• Fix the issue that prometheus doesn't collect drainer metrics (#1012, @SE-Bin)

Improvements

• Optimize the scene that the display command waits for a long time ,because the status cannot be obtained in time when network problems occur.(#986, @9547)
• Cluster and dm component support version input without leading 'v' (#1009, @AstroProfundis)
• When a user try to clean logs with the command tiup cluster clean --logs,add a warning to explain that we will stop the cluster before clean logs (#1029, @lucklove)

v1.3.0

New Features

• Modify TiFlash's query memory limit from 10GB to 0(unlimited) in playground cluster (#907, @LittleFall)
• Import configuration into topology meta when migrating a cluster from Ansible (#766, @yuzhibotao)
  • Before, we stored imported ansible config in ansible-imported-configs which is hidden for users, in this release, we merge the configs into meta.yaml so that the user can see the config with the command tiup cluster edit
• Enhance the tiup mirror command (#860, @lucklove)
  • Support merge two or more mirrors into one
  • Support publish component to local mirror besides remote mirror
  • Support add component owner to local mirror
• Partially support deploy cluster with hostname besides ip address (EXPERIMENTAL) (#948,#949, @fln)
  • Not usable for production, as there would be issue if a hostname resolves to a new IP address after deployment
• Support setting custom timeout for waiting instances up in playground-cluster (#968, @unbyte)
• Support check and disable THP in tiup cluster check (#964, @anywhy)
• Support sign remote manifest and rotate root.json (#967, @lucklove)

Fixes

• Fixed the issue that the public key created by TiUP was not removed after the cluster was destroyed (#910, @9547)
• Fix the issue that user defined grafana username and password not imported from tidb-ansible cluster correctly (#937, @AstroProfundis)
• Fix the issue that playground cluster not quiting components with correct order: TiDB -> TiKV -> PD (#933, @unbyte)
• Fix the issue that TiKV reports wrong advertise address when --status-addr is set to a wildcard address like 0.0.0.0 (#951, @lucklove)
• Fix the issue that Prometheus doesn't reload target after scale-in action (#958, @9547)
• Fix the issue that the config file for TiFlash missing in playground cluster (#969, @unbyte)
• Fix Tilfash startup failed without stderr output when numa is enabled but numactl cannot be found (#984, @lucklove)
• Fix the issue that the deployment environment fail to copy config file when zsh is configured (#982, @9547)

Improvements

• Enable memory buddyinfo monitoring on node_exporter to collect exposes statistics of memory fragments (#904, @9547)
• Move error logs dumped by tiup-dm and tiup-cluster to ${TIUP_HOME}/logs (#908, @9547)
• Allow run pure TiKV (without TiDB) cluster in playground cluster (#926, @sticnarf)
• Add confirm stage for upgrade action (#963, @Win-Man)
• Omit debug log from console output in tiup-cluster (#977, @AstroProfundis)
• Prompt list of paths to be deleted before processing in the clean action of tiup-cluster (#981, #993, @AstroProfundis)
• Make error message of monitor port conflict more readable (#966, @JaySon-Huang)

v1.2.5

Fixes

• Fix the issue that can't operate the cluster which have tispark workers without tispark master (#924, @AstroProfundis)
  • Root cause: once the tispark master been removed from the cluster, any later action will be reject by TiUP
  • Fix: make it possible for broken clusters to fix no tispark master error by scaling out a new tispark master node
• Fix the issue that it report pump node id not found while drainer node id not found (#925, @lucklove)

Improvements

• Support deploy TiFlash on multi-disks with "storage" configurations since v4.0.9 (#931, #938, @JaySon-Huang)
• Check duplicated pd_servers.name in the topology before truly deploy the cluster (#922, @anywhy)

v1.2.4

Fixes

• Fix the issue that Pump & Drainer has different node id between tidb-ansible and TiUP (#903, @lucklove)
  • For the cluster imported from tidb-ansible, if the pump or drainer is restarted, it will start with a new node id
  • Risk of this issue: binlog may not work correctly after restart pump or drainer
• Fix the issue that audit log may get lost in some special case (#879, #882, @9547)
  • If the user execute two commands one follows the other, and the second one quit in 1 second, the audit log of the first command will be overwirten by the second one
  • Risk caused by this issue: some audit logs may get lost in above case
• Fix the issue that new component deployed with tiup cluster scale-out doesn't auto start when rebooting (#905, @9547)
  • Risk caused by this issue: the cluster may be unavailable after rebooting
• Fix the issue that data directory of tiflash is not deleted if multiple data directories are specified (#871, @9547)
• Fix the issue that node_exporter and blackbox_exporter not cleaned up after scale-in all instances on specified host (#857, @9547)
• Fix the issue that the patch command will fail when try to patch dm cluster (#884, @lucklove)
• Fix the issue that the bench component report Error 1105: client has multi-statement capability disabled (#887, @mahjonp)
• Fix the issue that the TiSpark node can't be upgraded (#901, @lucklove)
• Fix the issue that tiup-playground can't start TiFlash with newest nightly PD (#902, @lucklove)

Improvements

• Ignore no tispark master error when listing clusters since the master node may be remove by scale-in --force (#920, @AstroProfundis)

v1.2.3

Fixes

• Fix misleading warning message in the display command (#869, @lucklove)

v1.2.1

Risk Events

A critical bug that introduced in V1.0.0 had been fixed in v1.0.8.
if the user want to scale in some TiKV nodes with the command tiup cluster scale-in with tiup-cluster, TiUP may delete TiKV nodes by mistake, causing the TiDB cluster data loss
The root cause:

1. while TiUP treats these TiKV nodes' state as tombstone by mistake, it would report an error that confuses the user.
2. Then the user would execute the command tiup cluster display to confirm the real state of the cluster, but the display command also displays these TiKV nodes are in tombstone state too;
3. what's worse, the display command will destroy tombstone nodes automatically, no user confirmation required. So these TiKV nodes were destroyed by mistake.

To prevent this, we introduce a more safe manual way to clean up tombstone nodes in this release.

Improvements

• Introduce a more safe way to cleanup tombstone nodes (#858, @lucklove)
  • When an user scale-in a TiKV server, it's data is not deleted until the user executes a display command, it's risky because there is no choice for user to confirm
  • We have add a prune command for the cleanup stage, the display command will not cleanup tombstone instance any more
• Skip auto-start the cluster before the scale-out action because there may be some damaged instance that can't be started (#848, @lucklove)
  • In this version, the user should make sure the cluster is working correctly by themselves before executing scale-out
• Introduce a more graceful way to check TiKV labels (#843, @lucklove)
  • Before this change, we check TiKV labels from the config files of TiKV and PD servers, however, servers imported from tidb-ansible deployment don't store latest labels in local config, this causes inaccurate label information
  • After this we will fetch PD and TiKV labels with PD api in display command

Fixes

• Fix the issue that there is datarace when concurrent save the same file (#836, @9547)
  • We found that while the cluster deployed with TLS supported, the ca.crt file was saved multi times in parallel, this may lead to the ca.crt file to be left empty
  • The influence of this issue is that the tiup client may not communicate with the cluster
• Fix the issue that files copied by TiUP may have different mode with origin files (#844, @lucklove)
• Fix the issue that the tiup script not updated after scale-in PD (#824, @9547)