Resolved Issues in 5.0.0-SP16

Prev Next

Install & Upgrade

  • ORION-155796: Enhanced handling of management virtual IP strings during upgrades to eliminate a flow that could result in upgrade failure because of inability to find the correct IP in the list of CNode interfaces.

  • ORION-152757: Resolved an issue that could cause a failure to activate one or more CNodes following an upgrade.

  • ORION-151644: Resolved an issue that caused a failure of SSD migration between two DNodes during VAST OS upgrade.

Cluster Expansion

  • ORION-173929: Resolved an issue that could cause multiple Failed running validation validate_box_chassis_serial: string indices must be integers alerts when starting the expansion procedure.

  • ORION-172211, ORION-176891: Resolved an issue that could cause cluster expansion to fail because of the assertion failed: (n_total_tags <= _max_n_tags) error.

  • ORION-172336, ORION-170731: Resolved an issue that could cause CNode containers to restart with the timeout expired for life_type=16,life_gen=15321536 (TRAVIS) error preceded by high write latencies during phase 2 (formatting new space) of cluster expansion.

  • ORION-163357: Updated post-expansion cleanup tasks to eliminate an issue that could cause a raised unexpected: IntegrityError('update or delete on table "vmsapp_host" violates foreign key constraint "vmsapp_dnode_host_id_b95a16d2_fk" on table "vmsapp_dnode") error when adding two DBoxes by means of the same DBox add task.

Networking

  • ORION-169912: Updated the InfiniBand monitoring routines so that VMS does not restart the OpenSM service every hour on all CNodes.

Element Store

  • ORION-173386: Introduced updates to ensure that running a VAST metadata analysis command does not cause the CNode container to restart with the timeout expired for life_type=0,life_gen=47996899 (INGEST_READ) error.

  • ORION-171110: Updated similarity-based data reduction to eliminate a flow that could cause CNode containers to restart with the “assertion failed: (gain_choosing_b != (18446744073709551615UL) && gain_choosing_d != (18446744073709551615UL)) invalid gain(s) given - gain_choosing_d=18446744073709551615 gain_choosing_b=0 - values will not be used to examine the box counter: 0 error.

  • ORION-142967: Introduced a number of improvements to eliminate a flow that could cause a repeated W_DR_DEFRAG had a suspension timeout alert on the cluster.

Quality of Service (QoS)

  • ORION-164889: Eliminated a QoS policy identification issue that could cause an qos policy remove returned an error: ObjectRemoveResultCode.NOT FOUND error when trying to delete a QoS policy.

Protocols

  • ORION-175600: Updated NFS access checking routines so that by default, an NFS client would be able to read a file that has an SMB read-only attribute (and resides on an NFS and SMB-enabled view). Prior to this change, the NFS client would get a permission deny error.

NFS

  • ORION-176475: Enhanced handling of NFS concurrent writes to improve performance when writing very large chunks of data to the cluster.

  • ORION-170152: Resolved an NFS locking issue that could cause CNode container restarts with the proc not found error accompanied with the NLM POST_DECODE maintenance deny list alert on clusters running VAST Cluster 4.7.0-SP18.

NFSv3

  • ORION-176931: Resolved an issue where after an NLM lock to read a file was requested by a large number of jobs, some jobs were able to obtain the lock while others were stuck in a waiting state.

  • ORION-157583: Resolved a cache buffer issue that could cause a CNode container to restart with the assertion failed: (is_reffed()) error after receiving a partial message from an NFSv3 client.

SMB

  • ORION-172731: Updated SMB directory handling routines to eliminate an issue that could cause CNode containers to restart with the assertion failed: (conv_len &lt; sizeof(file_name_glob_expression)) error while processing SMB workloads.

  • ORION-158199: Enhanced implementation of SMB Server-Side Copy to eliminate a flow that could cause an SMB client to fail back to regular copying (by reading and writing in chunks) instead of taking advantage of server-side copying.

VAST Database

  • ORION-174201: Resolved an issue that could cause CNode containers to restart with the assertion failed: (next(index) == INVALID_INDEX) error during processing of SMB requests.

Replication

  • ORION-177447: Added more validations to ensure that the name of the target bucket on an S3 replication peer does not exceed the maximum allowable length of 63 characters.

  • ORION-173194: Updated S3 bucket modification routines to eliminate an issue that could cause the assertion failed: ((arg_len) < (BUFF_SIZE)) (65 < 65) argument string too long error.

  • ORION-172303: Introduced changes into the logic of replication stream scheduling to prevent streams from being stuck following a stream failure.

  • ORION-165768, ORION-168095, ORION-176735: Improved handling of orphaned SIDs in cache during replication to eliminate a flow that could cause a local vid was translated to SidVal which doesn't appear in remote udb_cache counter: 0 alert on the replication target cluster and an ESTORE:NATIVE_REPLICATION deny list alert on the source cluster.

  • ORION-158587, ORION-139028: Resolved an issue that caused an OSError: [Errno 18] Invalid cross-device link error when trying to move files and directories within a protected path at the time when a quota was being created.

  • ORION-153209: Resolved an issue that caused a NATIVE_REPLICATION deny list alert on one of cluster’s CNodes due to the assertion failed: (is_gss_stream) Got STALE_HANDLE after handle translation!? error.

  • ORION-151748: Improved handling of a condition where a replication destination cluster exceeds its quota limit so that this condition does not cause a NATIVE_REPLICATION deny list on the source cluster.

Authentication & Authorization

  • ORION-151560: Updated the mechanism of querying a user in Active Directory so that VAST Cluster now also performs a query for parent groups of the user’s primary group SID. Prior to this change, parent groups were not queried, resulting in access deny errors in case the permissions were provided through a parent group.

  • ORION-151045: Updated VAST LDAP client to resolve an issue that could cause TLS certificate verification failure when using TLS with a CA-signed certificate.

  • ORION-145155: Enhanced handling of user identifiers when creating or modifying a quota to eliminate a flow that could cause a user not found error on a quota update attempt with NFS local/RPC authentication.

VMS

  • ORION-102961: Data that is protected by multiple protected paths is now included in capacity reports.

VAST Web UI

  • ORION-154967: Updated the logic behind the User Accounting tab in quota settings (Element Store -> Quotas -> choose to view or edit a quota -> go to User Accounting tab) to avoid displaying entries with an Identifier of 0 (zero) and N/A values in other fields in case the user has been removed from the authentication and authorization provider.

  • ORION-148221: Updated the logic behind the Tenant ID column in the VAST Catalog page (DataBase -> VAST Catalog) to avoid displaying negative values when there is only one tenant.

VAST CLI

  • ORION-161804: Updated the logic behind the vastcatalogconfig delete command to prevent getting a VastcatalogconfigApi: object has not attribute “vastcatalogconfig_list” error when attempting to delete a VAST Catalog configuration via VAST CLI.

  • ORION-142808: Added the value of UNKNOWN to the list of valid values suggested for the --state option on the ssd list and nvram list commands.

Platform & Control

  • ORION-164941: Enhanced DBox HA procedures to prevent a flow where disconnection of a splitter cable could cause a DBox failure.

  • ORION-164894: Resolved an issue that could cause CNode container restarts during an OS upgrade with the assertion failed: (success) Failed read locking an RWspinlock for 6000000 times error.

  • ORION-153935: Updated content block reindexing routines to prevent a flow that could cause unexpectedly high metadata usage on the cluster.

  • ORION-137929: Resolved an issue where a DNode fault isolation and recovery procedure did not succeed due to the assertion failed: (node->is_deactivation_safe()) node=DNodeType:<dnode> isn't safe for deactivation error.

Call Home & Support

  • ORION-154707: Improved the logic behind the Max upload bandwidth field in the call home settings dialog (Call Home -> General Setup) so that specifying a value in this field does not cause an unrecognized arguments error when trying to upload the bundle.