Releases: noobaa/noobaa-core
v5.3 'Bomberman'
Bomberman
Version Info:
- Git Branch: 5.3
- Label: v5.3.1
- Docker hub image:
Additions:
- Upgrade to Node@12
- Update NPM Dependencies
- Cleanup deprecated NAN code
- IBM Cloud Resource and NS Resource Support
- Separate endpoints to run stateless in an auto-scaling deployment
- Removal of endpoints from the core pod
- Removal of endpoints from agents (pv pools)
- UI Updates in the pv pools section
- New UI section for endpoints
- Add auto-scaler
- Support S3 BucketPolicy API
- Performance Improvements on Overwrites
- Performance Improvements on Object Deletions (New BG which reclaims the deleted objects)
- Added API to update BucketClass (=> Placement Policy), will update all buckets with the given policy to the new one
- Addition of k8s pools to Prometheus exported metrics
Internal:
Fixes:
Use yasm RPM in the Docker build #5916
read_bucket_sdk_info is set the request as anonymous #5838
Error in creating pv backend store #5805
A lot of RPC warnings "GOT RESPONSE BUT REQUEST NOT PENDING" probably on test_network_perf #5761
k8s pool status is stuck on initializing when a persistent volume cannot be allocated #5758
Add bucket usage to the Prometheus Metrics (OBC and OB) #5755
Account Default Resource should change only when the first pool is Healthy. #5694
Add bucket_api.create_bucket_for_claim() #5676
_check_disk_space in server_monitor is not relevant #5614
Tables header does not behave as it should on scroll #5432
Chunked uploads choke endpoint process (may crash other services as well) - Should pass CONTAINER_MEM_REQUEST #3495
v5.2 'Arkanoid'
Version Info:
- Git Branch: 5.2
- Label: v5.2.13
- Docker hub image: noobaa/noobaa-core:5.2.13
Additions:
- Performance fixes for cloud resource R/W flows
- S3 Website support
- Support deletion of a bucket with objects, they will be reclaimed in the background
- Support adding tiers to a bucket via the UI
- Replace NooBaa's cert mechanism to use a mounted k8s secrets
- Discover external addresses service by OpenShift routes (if exists)
- UI paginator component replaced
- Add links to entities in the audit log
Internal:
- SSL Cert & Events log moved to new UI arch
- Add ability to override config.js values via ENV variables
- Cleanup startup errors in logs
- Risky fault tolerance will not happen even if one replica only (bucket status)
- Update server min requirements alert thresholds
Fixes:
Set Debug Mode does not stay the configured time (10m) and goes back to 0 shortly after setting #5843
when ipify.org is blocked it causes all resources to be offline #5815
Generated SSL certificate error NET::ERR_CERT_REVOKED #5791
k8s pool did not update image after brain upgrade #5772
User without permissions to buckets can delete/edit the buckets using the Management Console #5769
Allowing to create an account with bucket creation false with internal as the default resource #5766
nb_native.init_rand_seed() stuck opening/reading from /dev/random #5745
block_store_client delegation low performance #5731
UI Should refresh after uploading a cert #5729
Performance: Optimize Small Object IO by using GET with small range instead of HEAD #5717
Create account with S3 access when no resources requires default_pool #5675
when the events are reach to the end it jumps up #5649
support delete bucket with objects - objects should be deleted in background #5460
Changing a multi-file function to a one-file zip shows an error message on execution file dropdown and prevent editing #5448
The graph in "Resources -> Cloud Resources -> -> Connected Buckets" should be centralized #5395
Agent has status 'Some ports might be blocked' without any reason #4156
Remove long queries from read_system #3401
v5.1
Version Info:
- Git Branch: 5.1
- Label: v5.1.2
- Docker hub image: noobaa/noobaa-core:5.1.2
Additions:
-
Deployment of NooBaa Agents as StatefulSets
-
Adjust the number of s3rvers spawn in containers
-
Changes to Agents management UI
-
UI Replace resources summary with a new design
-
UI Replace buckets summary with a new design
-
UI Merge pools tab and cloud resource tab into storage resource tab
-
UI Reimplement my connection tab with new design
-
UI genral fixes
Internal:
- Changes to agents flow - only support k8s agents
Fixes:
Allowing to delete last bucket #5673
Operator account holds the first resource as default resource and the resource cannot be deleted #5651
Missing help information in edit pool configuration modal #5648
Automated error handling for k8s pool scaling (and creation) #5640
Create system returns operator token that has no system or role in auth #5635
After create system, FE fails with error: missing required filed authorizedBy #5580
when deleting a pool that is referenced by a function read_system gets broken #5520
Tiering | migration between tier 1 to tier 2 is not happening. #5230
Delete host crashes the UI in some cases. #5211
When we are deleting deactivated node the status does not change to deleting node #5192
get_cloud_services_stats - does not return empty stats for configured services which where not used yet #5121
Node stays in migrating state- Node shows stored part although the data is not there anymore #4651
When delete host is finished we are not getting a popup message #4166
After deleting hosts we can see that there are activities even thought the host were removed from the UI #4164
Uploading file to system-internal-storage returns 403 AccessDenied when pool has no space #3848
Deactivated nodes should not be under the issue tab in the node filter #3088
Idle system with many buckets runs system_store.make_changes every second #2846
Nodes stuck in migrating after moving nodes between pools and retuning to the original pool #2728
Need to be able to deactivate/activate nodes from the pools list. #1904
Alert when moving node between pool if node has noobaa data on it #1894
v5.0
Version Info:
- Git Branch: 5.0
- Label: v5.0.1
Additions:
- Kubernetes SSO
- Split mongodb into its own container within the pod
- Remove the need for core container PV
- Upgrade process supports the new pod architecture
- Removal of platform-specific mgmt UI
- Prometheus metrics
Internal:
- Build process fixes - more efficient layers
Fixes:
lambda triggers fail because _make_aws_config in function_server uses the wrong endpoint for S3 access #5567
Full Changelog: v5.0.0...v5.0.1
v4.0
Version Info:
- Git Branch: 4.0
- Label: v4.0.2
Can't upgrade from < 4.0
Additions:
- UI Facelift to Patternfly 4
- Prometheus Metrics
Internal:
- Auto-detect Server's Connectivity (public IP / DNS)
- S3-SSE Support
- Tiering/Rebuild Refactor and Fixes
v3.0
Deprecated Version
Version Info:
- Git Branch: 3.0
- Label:
- Upgrade Package:
- OVA:
Can't upgrade from < 3.0
Additions:
- Resend Activation Code
- Block system config operations based on the platform
- Lambda screen rework and move to new arch
- Analytics view
- Tiering & Mapper Refactor
- Unmanaged Upgrade path
- Initial Prometheus integration
- Kubernetes Support
- Install NooBaa Server and Agents
- Operator - Deploy NooBaa, Upgrade
- Upgrade agents in kubernetes
- S3 Tagging
- NS Bucket Triggers
- READ event triggers
- Triggers Tab in Function Page
Internal:
- FE Refactor most of the leftover modals / areas into the new arch
- Move to Node.JS 10 & Update NPM dependencies (See #5274)
- Fixed Lambda creation logic for users with a non-valid default resource
- Move to work with ip instead of network (first_install_wizard, server_monitor)
- Increase S3 Coverage
- Sensitive data wrapper
v2.10
Version Info:
- Git Branch: 2.10
- Label: v2.10.4
- Upgrade Package: https://storage.googleapis.com/noobaa-releases/2.10.4/noobaa-NVA-2.10.4-8efd2e7.tar.gz
- OVA: https://storage.googleapis.com/noobaa-releases/2.10.4/NooBaa-2.10.4-8efd2e7.ova
Additions:
Internal:
Closed Issues on this release:
(Automatically generated by comparing issue.closed_at with releases.published_at dates)
(#5259) New OVA - create system doesn't update date/time automatically -- Customer issue
(#5233) Bucket page> Object tab in low resolution (minimum res of 1024) looks bad when versioning is on
(#5231) Cloud Services Consumption charts: New system without any writes shows many writes and and values for egress
(#5162) Phonehome -Cloud usage statistics is missing
(#5153) Data unavailability - Uploaded zip file is corrupted while the source file is fine. -- Customer issue
(#5151) upgrade in a loop during "Validating Package" (get to 95% and zero in a loop) -- Customer issue
(#5149) Mongo backoff is not clear when trying to troubleshoot / manually fix -- Customer issue
(#5140) gulp-inject-version has Proprietary license - should remove from Frontend deps
(#5111) S3 endpoint counter in the resource table is wrong -- Customer issue
(#5094) Disconnected cluster server status summary correction
(#4965) Delete of nodes takes over a day -- Customer issue
Backport of the following issues:
Fix AMZ date parsing
Fix unintended false value in XML reply
Adding support for chunked upload
Fix ObjectMDCache issue
read_bucket - only count objects in read_system
Build Code (Jenkins job changes)
v2.9
Version Info:
- Git Branch: 2.9
- Label: v2.9.3
- Upgrade Package: https://storage.googleapis.com/noobaa-releases/2.9.3/noobaa-NVA-2.9.3-b35c3ba.tar.gz
- OVA: https://storage.googleapis.com/noobaa-releases/2.9.3/NooBaa-2.9.3-b35c3ba.ova
Additions:
- Bucket usage breakdown in pool view (associated buckets)
- Regions assignment (and cloud resource page as a part of it)
- VMWare tools installtion
- IP collision detection sticky
Internal:
- BE data collections for analytics
Closed Issues on this release:
(Automatically generated by comparing issue.closed_at with releases.published_at dates)
(#5122) Pure storage resource has read/write problems -- Customer reported
(#5103) After a few days, all nodes became offline -- Customer reported
(#5091) S3 Endpoint counter shows 0 in FE on pools page when I have endpoints on agents
(#5086) Windows 2008 agent - FW detection fails
(#5084) Cluster - After bringing back the original primary (manually set higher priority) - all nodes are down -- Customer reported
(#5083) Windows endpoint is faulty after killing the primary core -- Customer reported
(#5066) S3 Endpoint on a node does not handle master change in cluster -- Customer reported
(#5059) on cluster upgrade - not all members were upgraded to mongodb 3.6 -- Customer reported
(#4894) Generic S3 compatible connection- We should provide a better error message in the Endpoint field
(#3231) A node was stuck on 100% CPU and 100% mem usage due to log file being locked/unavailable for access (HC:1) -- Customer reported
v2.8
Version Info:
- Git Branch: 2.8
- Label: v2.8.2
- Upgrade Package: https://storage.googleapis.com/noobaa-releases/2.8.1/noobaa-NVA-2.8.1-9587959.tar.gz
- OVA: https://storage.googleapis.com/noobaa-releases/2.8.1/NooBaa-2.8.1-9587959.ova
Additions:
- When writing via an endpoint, prefer writing to a pool which the endpoint resides in (provided its part of the placement policy, otherwise, keep old behaviour)
- S3 Versioning
- Pure FlashBlade resource type
Internal:
- Adding UserAgent to GCP usage report
- Update MongoDB to 3.6
- Update NVM and NPM on server
- Send first phone home report immediately after system creation
- Regions support, BE only
Closed Issues on this release:
(Automatically generated by comparing issue.closed_at with releases.published_at dates)
(#5027) Versions information is wrong (based on S3 browser failure) -- Customer reported
(#5010) Object with versions returns the old versions multiple times
(#5009) Can't preview old versions of an objects
(#4991) Can't delete buckets
(#4985) Installation on ubuntu 16 / 32 bit ends fine, but doesn't start the node nor show error -- Customer reported
(#4980) Delete connection is not working from the UI (pressing on it will only send read system)
(#4973) Tooltip position should be according to screen size and content length. They should not be trimmed.
(#4970) Discovery of new drives on node assigns them to irrelevant pool
(#4967) Configuration unsaved changes asterisk indication continues to be shown after changes are saved
(#4965) Delete of nodes takes over a day -- Customer reported
(#4953) Failed to create second bucket & first namespace bucket [cluster] -- Customer reported
(#4948) we should not allow two connections with the same name per same user
(#4940) If there are two connections with the same name (different types) - pressing delete in the ui affects both
(#4927) We allow creation of a connection to azure over a premium account storage - should fail
(#4923) Upgrade 2.6.3-c63dc65 to 2.7.268e122e stuck on 95% (phase II) - Failed error handling on supervisor update -- Customer reported
(#4881) the system is rebuilding all the data into the spillover when we reach out of space.
(#4869) in some cases spillback can take a lot of time
(#4865) DNS name configuration shows an error message before the user even entered any value
(#4849) After a long time a server is running supervisorctl fails with unix:///tmp/supervisor.sock no such file -- Customer reported
(#4808) Vertical scroll in edit data placement resources table
(#4723) blocks are not evicted from spillover when all of the policy storage is full
(#4576) When clicking enter key when the first checkbox has focus in a table, all checkboxes are selected
(#4398) Cannot connect application to a node endpoint according to the provided IP (azure internal IP) in the UI
(#4348) Gap - Missing resource icons in object block table
(#4311) Alerts lines are getting different after marking as read
(#4049) File upload failure tooltip flickers when hovers
(#4040) Status tooltips do not fit on small resolutions
(#3697) New FE arch - no handling for "live" params to state bound view models
(#3483) There is no sort in Storage & drives page
(#3473) Even when node's storage is disabled- ״used by noobaa״ in pie chart stays the same
(#3149) complete_object_upload does not compare etag and number of parts (on multipart upload)
v2.7 Pure Edition
Version Info:
- Git Branch: 2.7-pure
- Label: v2.7.5
- Upgrade Package: https://storage.googleapis.com/noobaa-releases/2.7.5%20(P%20Version)/noobaa-NVA-2.7.5-146ddec.tar.gz
- OVA: https://storage.googleapis.com/noobaa-releases/2.7.5%20(P%20Version)/NooBaa-2.7.5-146ddec.ova
Special Version for Pure, Based on 2.7.3 with both experimental flags turned on:
1. Don't use object metadata over s3 compatible resources
2. Delegate credentials to endpoints instead of using signed URLs
(#4832) NameSpace | trying to delete a file that is on the read only resource we get an internal error.
(#4407) NameSpace | Edit file does not delete the original file when rewriting it to a new bucket