Changes for Storage 9.4

Owned by Rumena Zlatkova (Deactivated)
Last updated: Apr 05, 2022 by Kyle Miller (Deactivated)
3 min read

New Features

This release improves configuration and troubleshooting and updates Swarm third-party components.

Elasticsearch Configuration

  • Swarm's Elasticsearch package now includes a script to streamline configuration of Elasticsearch servers. The script (configure_elasticsearch_with_swarm_search.py) makes needed changes to the elasticsearch.yml, logging.yml, elasticsearch logrotate.d, and elasticsearch systemd override files. (SWAR-7385)

  • The Metrics Curator now validates Swarm cluster names against existing search indexes to detect typos and other configuration problems; to bypass the validation, use the -v (--valid) option. (SWAR-7314)

Logging and Messaging

  • Error messaging for Metrics Curator has been expanded to clarify and cover more conditions. (SWAR-7314)

  • Logging and console errors alert when a replication feed is blocked and the cause of a problem. (SWAR-7547)

  • Swarm's dmesg dumps now include timestamps consistently across all interfaces, to help with diagnostics. (SWAR-7203)

OSS Updates — Swarm is now based on the Debian 8 (jessie) distribution, which brings numerous operating system bug fixes and improvements (SWAR-7519). In addition, these drivers and services have new versions:

  • Updated the Linux Kernel to 4.4.84, updated Linux Firmware to 1.167, and updated following Intel drivers: ixgbe 5.2.3, igb 5.3.5.10, ixgbevf 4.2.1, and e1000e 3.3.5.10. (SWAR-7605)

  • Updated Intel network drivers i40e to 2.1.26 and i40evf to 3.0.8 to support testing with newer Ethernet adapters. (SWAR-7643)

  • Updated the version of NTP to 4.2.8-p10 for security improvements. (SWAR-7639)

Upgrade Impacts

These items are changes to the product function that may require operational or development changes for integrated applications.

Impacts for 9.4

  • Upgrading from 9.2.0 or earlier — Because Swarm no longer generates the deprecated header Expect: Content-MD5, rolling upgrades in and remote replication with clusters running 9.2.0 or earlier are not supported. Recommended: Upgrade to 9.3.2 first, then upgrade to 9.4. (SWAR-7742)

  • EC and rolling upgrades — Because of changes to EC manifests in 9.4, sites performing EC writes by APPEND and Multipart APPEND (unlikely and rare) should contact DataCore Support before performing a rolling upgrade. Note, too, that the HP may report errors replicating EC manifests during the upgrade, but data integrity is not affected, and the errors stop when the upgrade is complete. (SWAR-7767)

  • Content Gateway 5.2.1+ required — If using Content Gateway, upgrade to version 5.2.1 or greater immediately after upgrading Swarm Storage.

  • KeepAliveInterval for Multipart — A new Swarm setting, scsp.keepAliveInterval, allows adjusting keep-alive messaging as needed for the SCSP client when performing multipart operations. It sets how many seconds to wait before sending successive chunked keep-alive bytes after a 202 Accepted response. (SWAR-7736)

  • Convert to EC with APPEND — Swarm now supports the ability of APPEND operations to convert objects from replication to erasure coding when requested via query arguments or lifepoint headers. (SWAR-7706)

Additional Changes

These items are other changes and improvements including those that come from testing and user feedback.

  • Resolved Issues

    • A failed multipart completion request mistakenly deletes the init stream, effectively canceling the multipart upload. (SWAR-7730)

    • Adding an SSD drive smaller than disk.minGB prevents other volumes on the chassis from mounting. (SWAR-7641)

    • Generation of dmesg or hwinfo dumps from the Swarm UI or legacy Admin Console occasionally hangs. (SWAR-7646)

    • hwinfo dumps from the Swarm serial port console can show errors. (SWAR-7635)

    • In a very full cluster, 507 Insufficient Storage and 503 Service Unavailable responses are returned on write and update requests. (SWAR-7574)

    • In the legacy Admin Console, creating or editing a replication feed with a domain restriction was rejected erroneously. (SWAR-7732)

    • Misuse of the recreatecid query argument as a method to move a bucket from one domain to another causes problems, so it is disallowed. (SWAR-7525)

    • No error message is generated if Swarm attempts to mount an encrypted volume without being configured with the necessary encryption key. (SWAR-7679)

    • On failed EC read requests, maximum recursion depth errors may be logged. (SWAR-7764)

    • Swarm nodes erroneously report "Low available memory is persisting". (SWAR-7637)

    • Using the preserve query argument on a COPY request changes the case of preserved headers. (SWAR-7628)

    • Very full nodes handling DELETE requests may fail with a 503 Service Unavailable response. (SWAR-7620)

© DataCore Software Corporation. · https://www.datacore.com · All rights reserved.