Aquesta és una versió anterior. La versió actual és la 2.0.7. HashiCorp Nomad 2.0.7 · Data de publicació:
2.0.2 forma part de la branca 2.0, que té suport fins al 30 d’abril del 2028. endoflife.date
Fitxers d’instal·lació 2.0.2
| Arquitectura | Tipus | Àmbit | Mida del fitxer | Baixa | SHA256 |
|---|---|---|---|---|---|
| 64 bits | ZIP → Portable | — | 64,4 MB | releases.hashicorp.com | VirusTotal |
L’enllaç i l’SHA256 provenen del manifest comunitari del Gestor de paquets del Windows. Convé comparar el hash del fitxer baixat abans d’executar-lo.
Notes de la versió
BREAKING CHANGES:
- logging: The allocation logs directory is bind-mounted read-only for task drivers that support with filesystem isolation [GH-27918]
SECURITY:
- dynamic host volumes: Prevent unintended code execution outside the plugin directory (CVE-2026-7474) [GH-27919]
- logging: Protect logging FIFO from symlink swap attacks (CVE-2026-6959) [GH-27918]
- sentinel: require sentinel-override ACL capability for overriding soft-mandatory policies on volumes
- ui: Upgraded Ember to 6.10 [GH-27674]
IMPROVEMENTS:
- acl: Downscoped AllowClientOp to node pool [GH-27718]
- api: Add "latest" flag for tagging the latest version of a job [GH-27764]
- build: Update Go toolchain to 1.26.3 [GH-27924]
- cli: Added retry for nomad job run monitoring [GH-27887]
- cli: Automatically expand nomad exec -it to -i -t [GH-27906]
- cli: job plan now propagates -hcl2-strict=false into the suggested nomad job run -check-index invocation when the user passed it on the plan command line [GH-23656]
- cli: add monitoring and verbose option to job dispatch [GH-27541]
- drivers: include volume RequestName within mount config information if available [GH-27710]
- scheduler: Add a configuration field for the number of nodes that the scheduler considers when spread or affinity is in use. This can improve scheduler performance for some cluster shapes. [GH-27650]
- server: RPC dial timeout is configurable [GH-27862]
- services: warn on job submit when job has services but no shutdown_delay [GH-27782]
BUG FIXES:
- acl: fix rpc permission denied error when using node_pool="all" [GH-27973]
- api: Fix a bug where the Create Job, Update Job, and Scale Job APIs could fail to respect EnforceIndex under concurrent requests [GH-27832]
- core: avoid setting job to dead while waiting for allocations to reschedule [GH-27852]
- csi: improve check of StagePublishBaseDir being subdirectory of MountDir [GH-27717]
- deployments: reset ProgressDeadline after pausing and do not fail while paused [GH-27804]
- drivers: kill plugin instance on dispense failure [GH-27711]
- plugins: Fixed a bug where plugin clients would continuously leak file descriptors when the agent was restarted [GH-27885]
- scheduler: Fixed a bug where preemption of allocations by tasks that require devices could incorrectly fail placement [GH-27880]
- ui: Fixed a bug where the evaluation detail panel would render improperly [GH-27987]
