Ceph : Issueshttps://tracker.ceph.com/https://tracker.ceph.com/favicon.ico2021-04-13T16:22:18ZCeph
Redmine Dashboard - Tasks #50340 (New): mgr/dashboard: CephFShttps://tracker.ceph.com/issues/503402021-04-13T16:22:18ZErnesto PuertaDashboard - Tasks #50335 (New): mgr/dashboard: Workflowshttps://tracker.ceph.com/issues/503352021-04-13T14:39:50ZErnesto PuertaDashboard - Feature #50334 (New): mgr/dashboard: cephfs-top and fsstats modulehttps://tracker.ceph.com/issues/503342021-04-13T14:22:28ZErnesto Puerta
<p>Same as for rbd-stats module and rbd-top-like Grafana dashboard (the one showing resource usage per RBD image), this would bring the equivalent CephFS functionality. We need to ensure that the fsstats metrics are stored in Prometheus (currently not) in the same way as rbd-stats are persisted via Prometheus exporter.</p> Dashboard - Feature #50333 (New): mgr/dashboard: CephFS mirroringhttps://tracker.ceph.com/issues/503332021-04-13T14:20:02ZErnesto Puerta
<p>Same as already done with RBD mirroring but for <a href="https://docs.ceph.com/en/latest/dev/cephfs-mirroring/" class="external">CephFS mirroring</a>.</p>
<p>This depends on <a class="external" href="https://tracker.ceph.com/issues/47261">https://tracker.ceph.com/issues/47261</a></p> Dashboard - Feature #50332 (New): mgr/dashboard: CephFS volumes, subvolume and subvolume groupshttps://tracker.ceph.com/issues/503322021-04-13T14:18:47ZErnesto Puerta
<ul>
<li>CephFS "volume" mgr module: <a class="external" href="https://docs.ceph.com/en/latest/cephfs/fs-volumes/">https://docs.ceph.com/en/latest/cephfs/fs-volumes/</a></li>
</ul> Dashboard - Feature #50331 (New): mgr/dashboard: CephFS scheduled snapshotshttps://tracker.ceph.com/issues/503312021-04-13T14:13:20ZErnesto Puerta
<ul>
<li><a href="https://docs.ceph.com/en/latest/cephfs/snap-schedule/" class="external">Cephfs Snapshot scheduler</a></li>
</ul> Dashboard - Cleanup #50330 (New): mgr/dashboard: rearrange RGW realms, zones, etc.https://tracker.ceph.com/issues/503302021-04-13T14:04:37ZErnesto Puerta
<ul>
<li>Replace "Object Gateway" menu with just "Object" [sage] or "Object storage"[Matt]</li>
<li>Change "selected object gateway" to "realm" [Sage]
* Display daemons, users, buckets by realm, not daemon (caveat: not all zones in a realm will belong to the same cluster)</li>
</ul> Dashboard - Cleanup #50329 (New): mgr/dashboard: provide URL to RGW daemonhttps://tracker.ceph.com/issues/503292021-04-13T14:02:00ZErnesto Puerta
<p>In order to make easier to consume the URL of an RGW daemon, it would be desirable to provide the URL of the daemon together with a copy-to-clibboard for direct usage.</p> Dashboard - Feature #50321 (New): mgr/dashboard: cephadm service spec schemashttps://tracker.ceph.com/issues/503212021-04-13T11:38:35ZErnesto Puerta
<p>Currently Dashboard contains hard-coded data structures of Cephadm internals (like Service Specs). This has led to some mismatches when Cephadm changes one of this internal specs. In order to avoid so, Service Specifications (and this might also be extended to other Cephadm internals) should be shared between Cephadm and Dashboard. As both components share code tree, the easiest way is to define abstract typed classes for Service Specs (if not yet) and import those from the Dashboard, where they would be inspected and the corresponding backend-frontend code would be generated dynamically.</p>
<p>This has already been done with other parts in Dashboard, like Cluster configuration, or Manager Modules (where the mgr_map is parsed to generate views of each mgr module OPTIONS).</p>
<p>It'll require that those classes are typed (python <code>typing</code>) and probably annotated with extra hints if they involve special validations (e.g.: max-min, etc).</p>
<p>As classes can be directly imported across projects, there's no need to explore schemas (like JSONSChema), which are only required from cross-language integrations (Python-Angular) or decoupled components (frontend/backend).</p> Dashboard - Bug #50319 (Resolved): mgr/dashboard: fix HAProxy (now called ingress)https://tracker.ceph.com/issues/503192021-04-13T11:29:10ZErnesto Puerta
<p>Right now HAProxy service creation is broken in Pacific: <a href="https://docs.ceph.com/en/latest/cephadm/rgw/#high-availability-service-for-rgw" class="external">the current Service Spec requires fields that the Dashboard doesn't allow to specify</a></p> Dashboard - Feature #50317 (New): mgr/dashboard: Basic/Advanced modehttps://tracker.ceph.com/issues/503172021-04-13T11:24:14ZErnesto Puerta
<p>Dashboard should allow users to toggle advanced/basic modes (default: basic). Advanced mode should be the Dashboard as we know it, while Basic mode would hide most non-essential input fields and would provide opinionated defaults for as much as possible.</p> Dashboard - Cleanup #50316 (New): mgr/dashboard: Edit EC profile: hide plugin lib directory https://tracker.ceph.com/issues/503162021-04-13T11:19:48ZErnesto Puerta
<p>Edit EC profile: hide plugin lib directory.</p> Dashboard - Cleanup #50315 (New): mgr/dashboard: refactor Configuration tablehttps://tracker.ceph.com/issues/503152021-04-13T11:18:34ZErnesto Puerta
<p>Simplify the current table: it contains too many rows that can be displayed in smarter ways (tooltips, icons, stylings).</p>
<p>This is the current one:<br /><img src="https://tracker.ceph.com/attachments/download/5451/configuration_page.png" alt="" /></p>
<p>The following is a proposal (only display setting name and current value/default):</p>
<p><img src="https://tracker.ceph.com/attachments/download/5452/configuration_page_proposal.png" alt="" /></p>
<p>In order to reduce the amount of default/basic settings displayed it should also be changed in the <a href="https://github.com/ceph/ceph/blob/master/src/common/options/global.yaml.in" title="now it's a YAML file" class="external">ceph core</a></p> Dashboard - Cleanup #50313 (New): mgr/dashboard: Do not rely on /dev/sdxhttps://tracker.ceph.com/issues/503132021-04-13T10:21:27ZErnesto Puerta
<p>Extend OSD creation wizard filters with path, but not the /dev/sdx, which can vary across reboots, but <a href="https://wiki.archlinux.org/index.php/Persistent_block_device_naming#by-uuid" class="external">relying on the UUID</a>.</p> Dashboard - Feature #50312 (New): mgr/dashboard: create OSD directly from device in Inventoryhttps://tracker.ceph.com/issues/503122021-04-13T10:15:26ZErnesto Puerta
<p>Add ability to create OSD directly from device at inventory page</p>