Release 3.0.79.25 * Do not clear replicaiton/export flags when unvaulted * Update fc driver to linux 6.6.66 * Extra memory checks during allocation * RHEL 9.5 build fix * Reduce memory allocations * Enable/disable file export for a VTL * SLES 15 SP6 build fixes Release 3.0.79.23 * Fix deduplication ratio drop when cloning tapes or migrating disks Release 3.0.79.22 * Rollback 'Reduce number of Anonymous Huge Pages', can cause daemon to crash Release 3.0.79.15 -> 3.0.79.21 * Fix excessive logging by iscsi daemon * Reduce number of Anonymous Huge Pages * Fix out of space conditions when sufficient disk space is available * Use dracut to rebuild initrd on OpenSUSE Release 3.0.79.15 * Fix performance regressions introduced in 3.0.79.14 Release 3.0.79.12 -> 3.0.79.14 * Alma 9.4 build fix * Fix memory leak in replication jobs with pool verification enabled * Performance improvements * Fix possible deadlock when an iSCSI connection is closed Release 3.0.79.12 * Fix memory leak in replication * Improve read/write performance with 4K logical sector disks Release 3.0.79.11 * Upgrade FC driver to 6.6.31 * Fix support for issuing SCSI UNMAP (discards) to disks Release 3.0.79.10 * Upgrade to lz4-1.9.4 * Fix coredev daemon crash when LTO encryption is enabled and tape block size > 256k * Fix database start failure in sqlrestore.sh * Fix ubuntu 20.x build failures * Switch to Linux 6.6 QLogic FC driver for all recent distributions Release 3.0.79 -> 3.0.79.9 * Use device definitions to configure a VTL. Device definitions allow for customized vendor and product strings * Improve write and erase performance for LTFS formatted tapes * Build fixes for Debian 12 * Switch to Linux 6.6 QLogic FC driver * Only listen on unix socket for postgresql server * Reduce memory allocations * Fix double free issue causing a kernel crash when socket connect fails * Copy iscsi chap settings when a drive is added to an existing VTL Release 3.0.78.1 * RHEL specific release. Add libpq-devel as a rpm install dependency Release 3.0.78 * Switch to distribution postgresql server. * Debian 12 and RHEL 9 support Release 3.0.77 * Ignore recovered errors while formatting physical tapes * Fix issues when FC write command data is not fully received * Fixes for exporting LTFS tapes * Initial support for secure boot * Show reported error when adding a disk fails Release 3.0.76 * Increase reserved memory mapped pages on startup * Switch to 5.15 fc driver * Use a separate event log for warnings and key events * Reduce memory usage by cached metadata * Reduce default running jobs to 3 from 32. * Handle multi-partition tapes during export Release 3.0.75 * Switch back to 4.9.x driver from 5.10 driver * Fix install on GCP kernels * Fix incorrect locking leading to out of memory conditions Release 3.0.74 * Reduce memory utilization * Import iscsiconf chap for upgrades from 3.0.58 Release 3.0.73 * Allow file and tape export scheduling for storage pools * Allow restarting services/system from GUI * Merge bugfixes from virtualization code Release 3.0.72 * Fix FC target-mode regressions, which caused device detection issues for certain hosts * Build fixes for kernel version 5.15 * Fix ietd daemon exiting on accept() socket error Release 3.0.71 * Fix deduplication issues since 3.0.60 leading to lower deduplication ratios and performance * Fix export issues of files to GCS * Ignore errors when trying to unload physical drive but there is no medium present * Fix issues with cleanup of directories in the cloud * Remove sock files under /var/run before start and after stop. Fixes service startup issues * Fix iscsi module load in RHEL 9 Release 3.0.70 * Fix regression in FC device detection * Create database user as a system user Release 3.0.69 * Fix fcint kernel module load failure Release 3.0.68 * Fix device detection by configuring FC ports after devices are configured. Release 3.0.67 * Use qlogic driver from 5.10.x kernel release * Delete remote VTL cartridges if source vcartridge is deleted Release 3.0.66 * Compress import/export job log files * Reduce log messages in import/export job log files * Allow both export and replication jobs for the same vcartridge * Fix detection of VMware disks Release 3.0.65 * Remove the need for a restart when VTL slot/ieport count is changed * Enable verification as the default for newly created pools * Enable auto-replication and auto-export for remote VTLs Release 3.0.64 * Cleanup ioctl warnings when loading deduplication tables * Log initiator name when PR registration is received * Include recent tables in sql dump/restore scripts * Pause replication/export jobs only when a disk is being replaced and not while a disk is being migrated * Fix iSCSI kernel crash when data for an invalid SCSI command is split into multiple PDUs * Cleanup kernel sockets on close, fixes kernel crash in newer kernels * Fix sockets 100% CPU usage when a socket has an error state Release 3.0.63 * Fix export errors with tape block size > 8MB * Fix export errors when EOD is encountered with out a trailing filemark * Allow vaulting of tapes from command line using vcconfig * Fix slow service shutdown down when RecvAddr is specified Release 3.0.62 * Include missing addchap.cgi * Handle invalide EVPD 0x83 reported by some libraries * Improve socket r/w performance * Fix VTL service startup failures * Support kernel 5.19 (ubuntu 20.04) * Add sysvinit-tools dependency for sles rpm * Increase vm.max_map_count to support systems upto 2TB of memory. * FC rule check bugfixes * Honor UseFreeSlot during unvault Release 3.0.61 * Set vm.overcommit_memory=1 on startup * Remove the need to set enable_encryption in conf for non-dedupe tapes * Fix crash when mmap fails * Build fixes for kernel 5.12+ Release 3.0.60 * Encryption fixes for tape block size <= 4096 bytes * Fix DB connection leak when quering a user * Add --vault option to vcconfig * Add logger debug for diag script * Speed up diagnostic collection * Fix daemon 100% cpu usage when querying storage pool disks Release 3.0.59 * Allow adding/removing a disk to/from Default pool from command line * Use ioctls for read from virtual drive instead of st driver * Fix reporting file identifer for read position cmd * Debian 11 build fixes * Use unix sockets for kernel->user notifications * Switch to /bin/bash instead of /bin/sh * Exec ldconfig before during service startup * Handle consecutive filemarks at the end of tape during a file/tape import * Allow importing file job to tape in specified slot * Allow configuration of hcache threshold * Change max_cmd_sn and add target alias in login response * Allow modifying both fc host and wwpn * Define process group, allow only users within the group to access daemon socket Release 3.0.58 * Fix medium_type reported for HP LTO volumes * Fix SLES SP2/SP3 builes * Remove default webadmin, let user create one * Clone vcartridges * Add support for Xen VBD disks * Build fixes for CentOS 8.5 Release 3.0.57 * Allow unvault of multiple vcartridges * Cancel pending jobs on tape mount * Fix double queuing of jobs during export jobs retry * Increase default log size to 32M * Fix high CPU usage during block allocations * Fix socket timeouts during replication over links with high latency Release 3.0.56 * Fix invalid sense key on write errors * Fix CentOS 8 stream install failures * Fix high CPU usage during block allocations * Fix disk detection issues with Netapp LUNs * Use multiple queues for FC command processing, improves multiple backup streams performance Release 3.0.55 * Fix detection of earlier configured disks. Some of them were not longer detected with 3.0.53/54 * Allow ':' in IQN names * Replication performance improvements * Reduce metadata writes during backups * Detect lvm volumes with '-' in volume group names Release 3.0.54 * Fix uncompression errors after a disk migration (2) * Improve replication performance for 4k dedupe blocks Release 3.0.53 * Fix double-free during iSCSI session reinstatement * Fix uncompressions errors after a disk migration * Allow _ and - in Host/Pool names * Fix build failures on Ubuntu 20.04 * Fix IBM i hosts incorrectly detecting LTO 1/2/3/4 drives * iSCSI interface performance improvements * Scan for Ceph RBD devices Release 3.0.52 * Improve FC driver handling for incomplete CTIO * Improve READ performance * Fix identification of MSFT virtual disks * Decrease deduplication table flush interval * Fix physical libraries going back to non-enabled state on startup * Improve iSCSI session id assignments Release 3.0.51 * Fix out-space errors when metadata loads are slow * Increase open file limit for VTL daemon * Extra logging for tape erase/delete Release 3.0.50 * Reduce memory allocation count * Fix auto-replication not working for some backup applications * Replication fixes * Handle issues with adapter restart when switching between initiator modes, fix softlock mtx_lock * Correct detection of LTO7/8 drives on OS/400 * Fix tape export to libraries without a barcode scanner Release 3.0.49 * Fix dbrecover to handle use db connection pools * Fix possible lockups when metadata load is slow * Add encryption support for non-dedupe pools * Handle invalid replication data when target pool verify is enabled * Detect SPECTRA T-Series physical libraries correctly * Use rcache when pool verify is enabled Release 3.0.48 * Fix replication when verification is enabled at destination storage pool * Fix physical tape library and drive detection on a rescan. * Skip temporary virtual devices created for import/export from showing up under "Physical Libraries" * Use read cache during replication * Command line /quadstorvtl/bin/vtrundiag for generating diagnostic logs * Add --rescan option to /quadstorvtl/bin/bdconfig to rescan for new physical disks/luns * Fix 'LoadError" seen for deleted tapes on a service restart. These tape should have been deleted from the DB Release 3.0.47 * Fix iSCSI and replication issues on CentOS 8 * Improve deduplication ratio of zeroed blocks for IBM iSeries BRMS * Fix general protection fault/traps on CentOS for random applications * Reduce backspacing during import/export * If VTL dedupe block size is set, use that during import Release 3.0.46 * Fix socket accept failures on 5.x kernels * PowerPC do not use Altivec or VSX assembly * Fix build failure on OpenSuse 15.2 * Fix a possible kernel crash when tape block allocation fails * Fix tape threshold checks for non-dedupe storage pools * Handle a rare case where tape block allocation fails even with free storage Release 3.0.45 * Fix decompression speeds and read performance * Utilize all available space on a disk * Fix blank FC host table due to bad format Release 3.0.44 * Use a DB connection pool * Fix service start failures due to change in load scripts * Fix startup failures on CentOS 8 due to SELinux * Stretch utilization of available disk space Release 3.0.43 * Fix another regression introduced by 3.0.41 where there is a possibility of a crash during erase or read Release 3.0.42 * Fix vcartrige load failures introduced by 3.0.41 Release 3.0.41 * Install udev rules for local device detection by Netbackup 8.2 * PowerPC64le packages for Debian 10 * FreeBSD 11.3 and 12.1 packages * Fix systemd start failures on CentOS 8.1 * User authentication for WebGUI * Build on Debian 10 * Map physical tape slots to virtual slots. Allows accessing physical tapes through VTL virtual drives Release 3.0.40 * Fix kernel crash when replacing a master disk * Fix replace failures when the source disk is a v1 metadata disk * LTO 8 emulation bugfixes * FreeBSD 12.1 support Release 3.0.39 * Make migration feature public * Fix install on RHEL 8.1 Release 3.0.38 * Fix high load average when trying to free metadata pages * Decrease metadata size for new disks * Fix kernel crashes during load or service shutdown due to extended_logging * Retain dedupe ratios during migration Release 3.0.37 * Fix timeouts when accessing GUI pages (add VTL, run diagnostics etc.) * Install on RHEL 8 and Ubuntu 1804, Sles 15 SP1 * Confirmation dialog when migrating or deleting a disk * Disable mixed mode drives in a VTL * Scan for NVMe devices * Do not use underscores in iqns Release 3.0.36 * Update migration code for faster replace of deduplication tables * Fix reduced usage of RAM for deduplication tables in some environments * Fix thread shutdown usage, possibly related to crash reports on service stop * Switch back to older metadata from when V5Tape=1 in quadstor.conf Fixes performance drop specific to an LTFS environment * Faster load of quadstor.conf * Add back support for file-backed storage pools Release 3.0.35 * Fix LTFS format failures for LTO-8 tapes * Fix LTFS reformat failures * Pause intermittently during ddtable loads to avoid incorrect reporting of soft-lockups during crc16 calculations on Debian Release 3.0.34 * Allow pausing auto-replication for a VTL * Speed up loading of deduplication tables * Wait till deduplication tables are completely in memory before enabling devices Release 3.0.33 * Additional log messages to track mount/unmount, PR reservations * Add support to use Linux 4.19 qla2xxx driver for RHEL/CentOS 7 * FreeBSD 11 support * Check threshold only for pools with AutoExport enabled and delete only vcartridges which have been exported to physical tape * Fix autoexport bug where the same vcartridge was exported to physical tape in a loop * Remove "Virtual Drives" from HTML GUI. No longer relevant * Skip additional page for selecting additional vdrives when creating a new VTL. This is no longer relevant since vdrives can be added after creating a VTL * Retry failed replication or tape export jobs after a delay * Replication performance improvement and configurable parameters * Wait for a tape job to cancel before completing a medium changer MOVE command. * Fix VTL service shutdown hang on Linux. Earlier the sock_shutdown() was not called correctly causing a socket thread to wait forever. This prevented the VTL service to shutdown and required a hard reset * Improve IBM iSeries compatibility for LTO 7 and 8 drives and 3584 library * Allow both upper and lower case when specifying a WWPN to fcconfing * Allow modifying Default pool properties from GUI/command line * Reload quadstor.conf via /quadstorvtl/bin/vtconfig --reload Release 3.0.32 * Handle fcconfig rules which apply to the entire VTL * Improve device detection from iSeries * Report barcode in 0x806 attribute - fixes warning reported by HP LTFS * Disk migration enhancements and bugfixes Release 3.0.31.1 * Unconfigured FC ports are shown as in Initiator mode whereas they are in Dual mode Release 3.0.31 * Fix device detection failures from iSeries * Fix physical storage GUI display issues * SLES SP3/SP4 and SLES 15 now supported * Allow fcconfig rules from GUI and changing role of HBA ports to target-only, initiator-only or dual * LTO 8 emulation * Improve SHA computation performance (upgrade to openssl 1.1.1a) * Reduce metadata write IO requests Release 3.0.30 * Reduce memory allocation for systems with < 16GB of RAM * Performance improvements * Fix possiblei but rare case of short reads * Improve migration code performance Release 3.0.29 * Improve support for iSeries * Reduce metadata writes - improved write performance * Initial disk migration support * Support upto 16MB -1 block size * Fix detection and associating physical drives withing a library Release 3.0.28 * Fix lower dedupe ratio regression * Fix decryption failures when encryption blocks are read * Add initial support for rclone * Allow load/unload of vcartridge from GUI * Fix "Data Size" calculations * Check for stale metadata when adding a new disk, requiring that the first 32MB of disk be cleared/zeroed in such cases * Fix for read failures in TSM * Update stats when a TAPE CAPACITY LOG PAGE is requested. Release 3.0.27 * Reduce ddtable syncs * Fix read cache size computations Release 3.0.26 * Fix build breakage on Ubuntu 16.04 * Fix possible list corruptions in FC driver * Fix reported soft lockups during ddtable loads * Fix wrong check for WORM media overwrite * Fix incorrect spacing over records when spacing backwards * Many replication fixes * Fix VTL daemon crash during replication of VTL with > 7 drives Release 3.0.25 * Fix incorrect 'Data Size' shown when duplicate blocks are found Release 3.0.24 * Show size of data on a vcartridge in the GUI * Fix possible out of memory conditions when erase threads are running in the background * Wait for atleast 12 minutes for ddtables to load during startup * Deduplication improvements (Netbackup, Networker) * Create maxiumum possible deduplication tables during disk initialization. Keeps deduplication tables sequential on disk Release 3.0.23 * Fix false warnings emitted when verification is enabled on a storage pool. Release 3.0.22 * Handle space positioning errors Release 3.0.21 * Linux 4.9.x build fixes * Fix fcconfig rules regression, VTL device earlier visible to hosts were no longer visible * Remove excessive syslog messages * Disallow new vcartridges when a disk is initializing Release 3.0.20 * Improve deduplication ratio for vm backups Release 3.0.19 * Fix install failures due to missing compilerdefs.h Release 3.0.18 * Fix for reads getting slower (and writes) over a period of time * Allow specifying incoming target port for fcconfig rules Release 3.0.17 * Misc. bug fixes * Fix install failures for recent kernels and OS distributions * Deduplication support for NetApp NDMP backups * Add viewing of log from System Page in GUI * Auto-replicate fixes * Skip EarlyWarning checks when importing from a file/tape Release 3.0.16 * Throttle writes when metadata reads are slower * Allow scanning for disk partitions with EnableDiskPartitions=1 in /quadstorvtl/etc/quadstor.conf Release 3.0.15 * Fix write failures due to false out-of-space conditions (2) Release 3.0.14 * Do not show disks with mounted partitions (Regression in 3.0.13) * Fix missing slab destroy on qla2xxx unload Release 3.0.13 * Correctly detect multipath disks Release 3.0.12 * Fix FC driver crash for 2.6.x kernels Release 3.0.11 * Updated FC driver * Fix write failures due to false out-of-space conditions Release 3.0.10 * Fix disk initialization failure introduced by 3.0.9 for pools with deduplicatiion enabled * Fix memory leaks due to list insertion without lock held Release 3.0.9 * Fix excess fragmentation of free blocks on physical storage Release 3.0.8 * Reduce metadata writes during erase operations. Speed up erase operations * Improve TSM dedupe ratios for large files and VM backups Release 3.0.7 * Fix replication kernel panics and performance improvements * Fix calculation of dedupe ratio for zeroed data * Reduce the number of parallel delete/erase operations to 2 Release 3.0.6 * Improve write IO performance. Fix bug where duplicate blocks were still being written to disk * Fix uninstall failures on RHEL 7. Stop vtl service before removing systemd unit file * Fix kernel panics on the destination of a replication job Release 3.0.5 * Fix backup failures due to dedupe table checksum mismatchs * Fix replication failures * Fix storage pool usage for replicated vcartridges * Allow deduplication for unsupported backup applications * Align metadata allocations on 4K block boundaries (4K sector disks) * Faster deletion and erase of tapes * Add systemd service unit file (RHEL7 and Debian 8) * Fix memory leaks in the destination of a replication op * Fix kernel panics when storage pool 'Verify' is enabled * Fix database recovery when drive count is more than 7 Release 3.0.4 * Fix for possible system crashes during reads due to an array index out of bounds * Fix AVOLTAG reporting. This fixes drive association for certain tape libraries in Data Protection Manager * Handle append-mode correctly for TSM Release 3.0.3 * Fix deduplication ratios when drive compression is disabled Release 3.0.2 * Fix RHEL/CentOS 7.2 installation failures Release 3.0.1 * Fix false tape checksum warnings when vtl service starts * Fix metadata overwrites when data inflight Release 3.0.0 * 3.0.0 release Release 2.4.51 * Auto-replicate and tape import/export bug fixes * Update FC driver to mainstream Linux 4.1.5 * Fix system crashes when handling task management functions in the FC driver * Fix CPU lockup under low physical storage conditions Release 2.4.50 * Sanity check device identifiers returned by physical disks, tapes and medium changers * Fix for intermittent failures of export/import operations Release 2.4.49 * Reduce random IO requests during a tape erase/delete. Decreases time taken for erase/delete * Improve backup write performance when physical disk space is low Release 2.4.48 * Disable deduplication when available disk space is low * Reduce the number of block hashes stored in the deduplication tables * Fix invalid log page length returned in response to a SCSI LOG SENSE command. This caused TSM server hangs/panics for older versions of lin_tape * Update FC driver to linux 3.14.34 * Reduce the number of VCartridge metadata blocks * Support for ALLOW_OVERWRITE command. * Check for Ubuntu 14.04 during install * Allow export to standalone tape drives from the GUI Release 2.4.47 * Allow modification of WORM flag for a storage pool and a vcartridge * SDLT specific sense which is required to determine the storage capacity remaining on tape was not returned with a REQUEST SENSE command. This is now fixed * Improve deduplication ratios for Netvault Unix and Netbackup * Fix vcartridge load errors after reformatting a multi-partition tape Release 2.4.46 * Netvault blanking of tapes did not reclaim used space. The issue occurs when a WRITE FILEMARKS command issued with the tape position at the beginning of tape/partition * Improvements to handling of READ ELEMENT STATUS command for large storage slot configurations. * Faster metadata writes. This also speeds up the time taken to delete a tape. Release 2.4.45 * Tape copy fixes for detecting EOD on source tape * TEST UNIT READY and LOAD UNLOAD commands didn't report correct sense data under certain conditions * iSCSI cmd response did not set correct residual value for some cases * Fix warnings reported when inserting entries into dedupe tables * Reinserting near full disks back into alloc lists when a certain threshold was free in certain cases lead to degraded write performance. This if now fixed Release 2.4.44 * Fix for devices sometimes not being detected over FC after a service restart * Fcconfig rule fixes where rule for a specify HBA applied to all vtls Release 2.4.43 * Minor optimization to handle multiple physical disks configured Fix for slow performance during writes Release 2.4.42 * Fix memory leaks after a write allocation failure * Handle a case where a WRITE FILEMARKS command is received and due to possible invalid metadata write, restores can fail Release 2.4.41 * For large disks (>= 2TB) with near full capacity, writes take too long due to block allocator scanning a large amount of metadata when the available RAM in the system is small This has been fixed using a more intelligent way of determining which disk to check for free blocks Release 2.4.40 * When block allocator metadata cache overflows try harder to free unused indexes. Fix for possible out-of-memory scenarios Release 2.4.39 * Reduce the number of times the free block thread scans for unalloced blocks Release 2.4.38 * Fix CPU soft lockup introduced by 2.4.37 Release 2.4.37 * Allow configuration of auto-replication schedule allowing for replication only during specified hours * Optimize block allocation giving priority for disks with higher free space Release 2.4.36 * Small optimization to check if deduplication tables need to be flushed before starting the sync thread * Deduplication tables read write optimizations Release 2.4.35 * Deduplication enchancements for backups from TSM and other applications * Allow addition of virtual drives for a VTL from the HTML GUI * Increase the maximum number of virtual drives per VTL to 64 * Allow modification of storage slots and i/e ports count using vtconfig utility Release 2.4.34 * Fix metadata checksum warnings for deduplication tables due to modification of deduplication pages while waiting for write IO to complete Release 2.4.33 * Fix regression introduced in 2.4.29 due to which flushing of deduplication tables to disk is significantly higher and impacts backups and restores * Reduce background IO during restores Release 2.4.32 * Urgent fix for deduplication tables sync thread in an infinite wait Release 2.4.31 * Fix for possible slow shutdown times due to sleep between dedupe tables sync * Fix for possible memcpy overflow of the write buffer for TSM backups * Deduplication ratio improvements for TSM backups * Reduce the number of new deduplication tables created, using table entries for data erased or deleted from a tape Release 2.4.30 * If a rewind,locate/space command is received after a series of reads, readahead pages/data may be freed before the disk read operations complete. This can lead to the pages being written to *after* they were freed and can lead to memory corruptions and system panics Release 2.4.29 * Sequential dedupe table read and writes wherever possible * Reduce the number of small sized I/O requests * Reduce the number of dedupe table writes during backups and restores * Additional check on metadata and boundary checks to prevent memory corruptions * Wait for a maximum of three minutes for the deduplication tables to load after which continue loading them asynchronously * Improve tape erase and deletion times further * Fix for possible cases where a backward spacing over a tape might fail Release 2.4.28 * In case a virtual tape is still in a drive during service stop/shutdown, after freeing memory associated with the tape we try to flush the buffers for the tape (although the buffers would have been flushed much buffer). This leads to usage of memory after it is freed and can potentially cause a crash Release 2.4.27 * Reduce the number of database operations during shutdown. With a large number of vcartridges configured, it has been observed that postgres reaches and stays at 100% cpu utilization * Reduce the number of deduplication tables flushed to dish during a tape erase operation * fcconfig no longer expects WWPNs to be lower case only * Wait for a limited amountof time for a task to be aborted. If the timeout is reached before the task could be aborted report an erro back to the initiator Release 2.4.26 * Performance improvements for parallel backup sessions * Using multiple drives for backups simultaneously led to the possibility of an infinite wait by one of the drives. This has been fixed Release 2.4.25 * Optimizations when receiving large blocks of data containing zeros * Wait for certain metadata to complete IO before progressing. There is a possibility of a soft lockup or a crash * Add additional assertion checks for page/memory references * Deleting or removing a disk from a non-dedupe pool resulted in a crash. Release 2.4.24 * Deduplication statistics and vcartridge usage reported for a disk can be lower than actual when replicated data is written to a vcartridge * VCartridge used size was different between the source vcartridge and the destination vcartridge after replication. This was because of an inconsistency in the disk space calculations for deduplicated and compressed data. Release 2.4.23 * Reduce time taken to delete a tape * Fix slow shutdown or infinite wait during quadstor shutdown * SCSI device reset improvements * Asynchronously load deduplication tables Release 2.4.22 * Fix possible crash by a command deletion by an ABORT command when the command is still in the write queue. * Limit sequential reads of the data deplication tables to 32 MB per pass. This whould limit the load in shared SAN environments * Warn in syslog if a write command takes more than 4 seconds Release 2.4.21 * For master disks (disks with deduplication tables) reserve atleast the space required for remaining data deduplication tables. Currently we reserve 1GB but that wouldn't be enoughf for large RAM configurations Release 2.4.20 * Minor replication issues. If auto-replication fails we should try later. * quadstorvtl service failed to shutdown cleanly if there were pending replication/import/export jobs * Limit the number of vcartridges that can be added at a time to 64 * Virtual devices may hang indefinitely when writes failed due to out of space conditions * Allow deletion of upto 64 vcartridges at a time from the GUI * Speed up initial load of data deduplication tables * Add initial set of command line utilites Release 2.4.19 * Expand deduplication support matrix * Verbose syslog messages during startup Release 2.4.18 * Support data deduplication for Veeam datasets Release 2.4.17 * Optimize data deduplication for Netvault datasets (spanning etc) * Flush pending writes before reporting TAPE CAPACITY log page. This would report the remaining space on a tape correctly * With auto-replication enabled, if a blank tape was mounted and then unmounted without any writes, replication should not be triggered for such tapes Release 2.4.16 * Support data deduplication for Netvault datasets Release 2.4.15 * Shutdown of quadstor service used to indefinitely wait if vcartridges are being deleted. This is fixed. * Add generation of diagnostics from HTML UI (System Page) to aid product support * Allow configuration of auto-export, auto-replicate and threshold for default pool * Trigger auto-replicate when a tape has been unmounted from a drive Release 2.4.14 * Fix soft lockups experienced with TSM backups. * Fix kernel crashes when writing blocks not aligned on a 4K boundary (For example a tape block size of 65535). TSM backups were the most affected * Fix dedupe ratio shown in the GUI. We weren't taking into account zero data blocks which are always deduplicated Release 2.4.13 * Additional check to prevent soft-lockups during deduplication scans * Disable deduplication checks if Logical Block Protection is enabled for the drive * Fix for data corruptions in cases where we wait for metadata load in order to determine if a block is a duplicate or unique * Support byte verification of possible duplicate blocks if enabled for the pool Release 2.4.12 * Improve READ/WRITE performance * Remove replication debugs * Fix READ after LOCATE commands * Single package installs instead of two package installs Release 2.4.11 * Fix for possible out of memory (OOM) conditions after many backups and/or restores * Fix for possible CPU lockup with deduplication enabled and certain backup datasets * Use SSSE3 optimized SHA-256 computations * Single package installations rather than two package (core and itf) installs * Build missing kernel modules automatically if kernel is upgraded * Support Fusion-IO and Virident Flash devices (/dev/fio and /dev/vgc) Release 2.4.10 * Physical tape import/export jobs can fail for the second job onwards for the same physical driver * Add support for application MAM attributes in tape import/export Release 2.4.9 * Write BUG fixed where blocks can be written out of order causing reads to fail Release 2.4.8 * Add, cancel and delete replication jobs from the HTML UI * When replicating compressed blocks, uncompress the blocks read before computing the hash. Earlier the hash computation was done on the compressed blocks, which is incorrect and can lead to incorrect deduplication on the remote VTL * Replication to new vcartridges for a standalone VDrive failed since during creation they were mounted on the remote VDrive. A fix has been added * MAM information from the source vcartrdige is now replicated to the destination vcartridge. * When starting a manual replication job, if the source address isn't specified but RecvAddr is present in quadstor.conf, use the RecvAddr as the source address * Fix numerous warnings generated when replicating compressed data Release 2.4.7 * Handle deadlocks when receving multiple fixed block writes (Reproducible with Data Protection Manager) Release 2.4.6 * Deduplication support for networker database and application backups Release 2.4.5 * TSM data deduplication support * Backup Exec data deduplication support * Data Protection Manager (DPM) data deduplication support Release 2.4.4 * FC LUN reporting improvements on a QUADStor VTL service restart * Data Protector Data Deduplication support * Fixed REPORT ELEMENT STATUS response. TSM could not inventory volumes nor associate drives with library due to a regression Release 2.4.3 * Support data deduplication for ArcServe backup sets * Report tape alert flags from 1 - 64. Fixes false warnings reported by Data Protector. * Revert back to 10 byte serial numbers for Quantum SDLT drives * Fix for READ crash after a REWIND because we did not free previous READ caches Release 2.4.2 * Increase application support for data deduplication * Fix page leaks with newer linux kernels * Reduce the number of slab allocations * Fixed NULL pointer dereference issue in infiniband code Release 2.4.1 * Fix kernel crash when an ERASE command is received for a blank tape * Fix kernel crash when reading data after a SPACE/LOCATE of a small range of blocks * Fix for READ errors after a SPACE/LOCATE of a small range of blocks * Space utilization reported should consider space reduction due to compression of data * Fix deduplication rations for Unix backups and backups over LAN * Fix out of memory conditions when reading/writing Release 2.4.0 * 2.4.0 release