TcaplusDB 3.62.1 Release Note

1. Feature List

  • Add: [Commercial Version] TcaplusDB to support periodic table truncating;
  • Add: [Commercial Version] TcaplusDB to support the number of shards of a single tcapsvr from 512 to 2000;
  • Add: [Commercial Version] To support the generation of automatic expansion and shrinkage plans according to the Shard dimension;
  • Add: [Commercial Version] Derivative tool's TEXT data source to support INT and other data types;
  • Add: [Commercial Version] Full-automatic tcapsvr slave rebuild and full-automatic detection trigger slave rebuild to support model upgrades;
  • Add: [Commercial Version] To support automatic management of accessing Polaris load balancing when non-Tencent Cloud scales tcaprestproxy
  • Add: [Commercial Version] To support Tcaprestproxy automatic expansion;
  • Add: [Commercial Version] To support local TOPN index (with tcapsvr version 3.62.0);
  • Add: [Tencent Cloud] Financialization transformation: the account on the whitelist is prohibited from providing services;
  • Add: [Tencent Cloud] TcaplusDB would freeze services by app when Tencent Cloud's arrears are frozen.
  • Add: [Commercial Version/Tencent Cloud] TcaplusDB - Check and filter resources that do not meet the version requirements in multiple types of transactions such as tcapsvr migration based on the set dimension;
  • Optimize: [Commercial Version] Access layer to support forced shrinkage of the shared Tcaproxy when the machine is down; the proxy can be migrated to other machines in case of machine failure.
  • Optimize: [Commercial Version] Tcapcenter to give early warning when tcapdb/tcapsvr/tcapproxy/restproxy resources are insufficient;
  • Optimize: [Commercial Version] Tcapsvr slave to support cold backup rebuild without downloading and redoing ulog, but directly recovering from the cold backup data, and then registering with the master before performing data synchronization
  • Optimize: [Tencent Cloud] It is necessary to add support for 9 & 11 & 30 network segments by default in Tencent Cloud mode, when the machine joines the system.
  • Optimize: [Tencent Cloud] CGI to recycle exclusive resources when deleting shared apps and roll back resources used when creating an exclusive set fails:
  • Optimize: [Commercial Version/Tencent Cloud] To support tcapsvr slave rebuild if the source machine hangs up which is in the process of lossless or lossy migration.
  • Optimize: [Commercial Version/Tencent Cloud] Tcapsvr Slave rebuild and rollback process - Optimization;
  • Optimize: [Commercial Version/Tencent Cloud] Tcaprestproxy security stop - Optimization: Tcapcenter refreshes the route before stopping;
  • Optimize: [Commercial Version/Tencent Cloud] Tcapcenter and Tcapproxy interaction performance when setting read shunt - Optimization;
  • Optimize: [Commercial Version/Tencent Cloud] Optimization: the issue that the number of fields in the table meta information may be inconsistent during physical migration;
  • Optimize: [Commercial Version/Tencent Cloud] Tcapdir Performance - Optimization: the issue that LoadDB timeout causes client authentication and service startup timeout;
  • Optimize: [Commercial Version/Tencent Cloud] To support the Tcapsvr master and slave switches during the index rebuild process.
  • Optimize: [Commercial Version/Tencent Cloud] TcaplusDB single set to support bus channel counts ranging from 3000 to 30000.

2. Bug List

  • Repair: the issue that the tcapdb cache synchronization is suspended when the shard is recycled by the shrinkage and migration transactions, which causes the shard recycling action to be intercepted;
  • Repair: the issue in which the page or transaction does not automatically determine whether there are any unfinished offline transactions, causing the offline transaction execution to recycle resources incorrectly before the machine is chosen and rebuilt as a new machine;
  • Repair: the issue that a small number of machines are occasionally missing the decrypt_tools and are not detected by the uploaded check action script, during the batch-executing machine initialization process;
  • Repair: incorrect logic of the Tcapsvr slave rebuild transaction script in processing ulog faults;
  • Repair: the issue that the Chinese string may get stuck when Tcapdb synchronizes data to es;
  • Repair: the issue that when the number of local indexes of a single table exceeds 8, the front-end is not blocked;
  • Repair: the issue of a data error after adding any value field to the generic table's existing value, and then migrate to another new machine; 
  • Repair: the issue that the button at the bottom cannot be clicked when the list is too long on the shard batch migration page;
  • Repair: the issue of adding pb table without protection for nested structure which may cause parsing burst stack;
  • Repair: the issue that deleting a table in Tencent Cloud will cause the route of the restproxy to be deleted;
  • Repair: the issue that if the version of shared Tcaproxy to be bound is less than that of TcapRestProxy when a client creates a zone, no Tcaproxy may be available because Tcaproxy is not added to the table.
  • Repair: the issue that the Shard could not be successfully recovered when a new Shard was allocated in a transaction waiting to be executed.
  • Repair: the issue that the transaction failed to suspend after changing the route in the master/slave transaction.
  • Repair: the issue that the DescribeClusters interface of Tencent Cloud CGI intercepts the data returned by Tencent Cloud;

results matching ""

    No results matching ""