Archive:Hints for beta testing

Note

IMPORTANT – READ CAREFULLY SEP sesam provides SEP sesam 5.0.0.1 Jaglion beta version as a test program. This means that 5.0.0.1 Jaglion beta is provided on an "as is" basis and without warranty! Beta version may contain errors and bugs and may cause problems that could lead to system failures and data loss. THEREFORE, SEP sesam BETA IS NOT INTENDED OR SUITED FOR PRODUCTIVE USE! By downloading, installing and using SEP sesam beta software, you acknowledge that you are aware of the beta release version status and agree that the use of the beta version is entirely at your risk. SEP AG excludes all liability for any claims, losses and damages arising out of use of beta version. Also excluded is the liability for used third-party products and for any data loss due to SEP sesam beta software-related system or hardware failures. If you do not agree to these terms, DO NOT INSTALL OR USE THIS BETA VERSION.

Testing SEP sesam 5.0.0.1 Jaglion beta

If you are interested in testing the beta version, please register as a beta tester by email to beta@sep.de. To learn about new features, check the Release Notes 5.0.0 Jaglion. Then read the following notes and hints that can help you to get started with your BETA test.

Known issues and limitations

	Note
	Antivirus programs may disrupt network communication and cause SEP sesam processes, such as backup and replication, to fail. One program that is known to cause SEP sesam processes to terminate is Sophos Firewall with IPS (Intrusion Prevention System) enabled. Make sure that there are no antivirus, firewall, IDS or IPS programs preventing interaction with SEP sesam.

5.0.0 Jaglion known issues:

Severity: SECURITY SEP sesam v. 4.4.3.70-5.0.0.4 – RCE vulnerability in Spring Framework running on JDK 9+ As reported by VMware (Spring Framework RCE, Early Announcement), two serious vulnerabilities have been found in Spring MVC and Spring WebFlux applications running on JDK 9+ that may lead to remote code execution (RCE). The specific exploit requires the application to run on Tomcat as a war deployment. To be vulnerable to the exploit, the application must be deployed as a war file in a standalone servlet container. The execution of an executable war file with an embedded container should not be affected by the vulnerability. However, as the nature of the vulnerability is more general, there may be other ways to exploit it. Even though SEP sesam does not ship any war files and is not executed as a servlet and is therefore not affected by this vulnerability, it uses Spring Core version 5.3.16 and will provide a fix in Jaglion V2 by integrating Spring Core version 5.3.18. For more details on the "Spring4Shell" vulnerability, see VMware vulnerability report CVE-2022-22965: Spring Framework RCE via Data Binding on JDK 9+. Resolution: Will be fixed with Jaglion V2. Newer sm_ui.jar files containing the fix are available on request for Beefalo and Jaglion V1. You can contact SEP support to get a newer version of sm_ui.jar. SEP sesam v. 5.0.0.x – (Windows) STPD automatically disables the HTTPS port on Windows (Applies only to Windows) The Sesam Transfer Protocol Server (STPD) automatically disables the HTTPS port if CPU instructions are missing due to unsupported AVX. Consequently, the TLS key and certificate cannot be created. Workaround : Make sure that you use CPUs with supported AVX versions (AVX, AVX2 or AVX-512) as well as FMA3 (Fused Multiply-Add 3-operand Form) or FMA4 (Fused Multiply-Add 4-operand Form). For the list of supported AVX versions, see Advanced Vector Extensions. Severity: MAJOR SEP sesam v. 5.0.0.2-5.0.0.x – Linux backup silently skips a directory if .nosbc file is part of backup source (Linux only) When backing up a backup source that contains one or more .nosbc files, a directory is silently skipped, but the backup is marked as successful. For example, if there is a backup source /data that contains /data/ds1/.nosbc, /data/ds1/pages/ and /data/ds1/trash/, then the pages/ or trash/ directories are not backed up. If a backup source does not contain a .nosbc file, all files are backed up successfully. Workaround: Since sbc incorrectly excludes a directory from backup if a file .nosbc is part of the backup source, use the -o noexcl switch in the backup options (backup task properties, tab Options -> Backup options) when backing up a backup source with a .nosbc file. Resolution: SEP is working on this issue and will provide a fix as soon as possible. SEP sesam v. 5.0.0.3-5.0.0.4 – Hyper-V restore of Hyper-V RCT backup fails with error When restoring a Hyper-V VM from a Hyper-V VM RTC backup under a new VM name, the restore fails with an error: Failed to read metadata file. Resolution: SEP is working on this issue and will provide a fix as soon as possible. SEP sesam v. 5.0.0.4 – SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects. This is due to an issue that the SharePoint Services Writer is missing from the writers list. Resolution: Will be fixed with the next version. Severity: NORMAL SEP sesam v. 5.0.0.4 – Replication between Si3 and Si3 NG does not work with a Windows RDS (incl. SEP sesam Server) (Windows only) Replication between Si3 and Si3 NG does not work with a Windows RDS (incl. SEP sesam Server) as sm_dedup_interface cannot read the message from the pipe correctly. The problem does not occur on Linux. Resolution: Will be fixed with the next version. If you experience this problem, SEP support can provide a hotfix. SEP sesam v. 5.0.0.1-5.0.0.4 – Check for completed Si3 -> Si3 NG replication not available In this case, replication is done by reading source-side backups, but no specific information is written to the Si3 replication log or displayed in the GUI. Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.x – Backup of SEP sesam client version 5.0.0 to SEP sesam Server version 4.4.3.x fails with "XBSA: Perform: curl error 55: Failed sending data to the peer" If package updates are installed on SEP sesam Clients during the regular update before SEP sesam Server is updated to the same version as the clients, the backup may fail with the message "XBSA: Perform: curl error 55: Failed sending data to the peer" Resolution: Update the SEP sesam Server to the same version installed on the clients. Note that you always have to update the SEP sesam Server first before updating the client software. SEP sesam v. 5.0.0.3-5.0.0.4 – Kopano backup on Windows fails with error Kopano backup on Windows fails with charmap' codec can't encode characters <char> in position <position>: character maps to <undefined>. The exception is caused by special characters such as '★' in the email subject. Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.3-5.0.0.4 – New Hyper-V RCT backups do not support single file restore (via VFS mount) As of 5.0.0 Jaglion, SEP sesam uses Resilient Change Tracking (RCT) by default for Hyper-V backups on Windows Server ≥ 2016 (Hyper-V VM ≥ 6.2). Single file restore (VFS mount) is currently only supported for VM Hyper-V backups without RCT option (old approach with VSS snapshots creating .avhdx files instead of the new Resilient Change Tracking). The new RCT backups have a 'RCT' attribute and the mount option is not shown. For older backups without the 'RCT' attribute, the mount option is available. Workaround: When restoring a Hyper-V RCT backup, perform a regular VM restore of the virtual machine as described in Restoring a Hyper-V virtual machine. Resolution: SEP development is working on a single file restore via mount for Hyper-V RTC backups and will implement it as soon as possible. SEP sesam v. 5.0.0.3-5.0.0.4 – Hyper-V restore fails, if the data store path from the original VM does not exist on the target server Hyper-V restore fails with error Warning: Cannot create item [D:\]: [87] WIN32 API error: 87 - The parameter is incorrect., if the data store path from the original VM does not exist on the target server, regardless of whether the default Hyper-V folder for VMs is used or an existing data store path is set. The original path is always used. Resolution: SEP is working on this issue and will provide a fix as soon as possible. SEP sesam v. 5.0.0.3-5.0.0.4 – (Debian only) Restore from RDX backup fails with error (Applies only to Debian): Restore from RDX backup fails with an error: Error while unpacking data. Bad archive. This problem is caused by incorrect segment parsing. Resolution: Will be fixed with version 5.0.0.5. If you experience this problem, SEP support can provide a hotfix. SEP sesam v. 5.0.0.3-5.0.0.4 – NDMP directory restore on NetApp 7 fails NDMP directory restore with seldir argument is not compatible with NetApp version ≤ 7, causing the restore to fail. Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.3-5.0.0.4 – Selective restore from NetApp Snap Store incorrectly restores all files and may overwrite existing files Selective restore from NetApp Snap Store incorrectly restores all files instead of the selected files and overwrites existing files when the "Overwrite" option is selected. Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.1-5.0.0.4 – Mounting Nutanix using migrated saveset fails Mount of Nutanix VD using migrated saveset fails with ERROR: sm_vfs could not be started: Cannot find matching saveset for LIS/LSL line: Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.3-5.0.0.4 – SEP sesam server shutdown hangs due to unreachable DB SEP sesam Server shutdown hangs when terminating SEP sesam processes because DB is no longer reachable. Resolution: Will be fixed with the next version. GUI issues We are experiencing some GUI issues that do not significantly affect the performance and functionality of SEP sesam. SEP is working to resolve these issues as quickly as possible. SEP sesam v. 5.0.0.3-5.0.0.4 – If a double quote exists in the annotations, tags or custom attributes of a VM, the VM cannot be browsed When a double quote appears in the VM annotation, tags or custom attributes, the VM cannot be browsed in ESX or vCenter due to broken parsing. Workaround: If you experience this problem, contact SEP support to get a newer version of sm_ui.jar. Resolution: Remove special characters in annotations, tags or custom attributes. The issue is addressed and will be fixed with the upcoming Service Pack 2 release. SEP sesam v. 4.4.3.84–5.0.0.4 – If a manual snapshot of the VM exists, the vSphere DIFF backup fails When a manual snapshot is created for the VM, the snapshot number is not processed correctly and the *.cbt file on the datamover is not created, causing the backup to fail. Workaround: If you experience this problem, contact SEP support to get a newer version of sbc_vadp.jar. Resolution: The issue is addressed and will be fixed with the upcoming Service Pack 2 release. SEP sesam v. 4.4.3.84–5.0.0.4 – When setting a backup source for a Microsoft SharePoint server backup by browsing, the backup source is invalid When browsing items of a Microsoft Sharepoint server to select a backup source, the created backup tasks are not valid because the check expects a leading slash ('/') that is not part of the Sharepoint backup source path. Workaround: Edit the backup task and correct the backup source manually by removing the '/SharePoint Server:' part from the source path. Resolution: The issue is addressed and will be fixed with the upcoming Service Pack 2 release. Severity: MINOR SEP sesam v. 5.0.0.3 – The command sm_dedup_interface dir full lists no items The sm_dedup_interface dir full command does not work, so the SI3 fsck report shows "Saveset not found on disk", even though saveset files exist. Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.4 – (Windows) A core dump occurs during SEP sesam backup on Windows (Applies only to Windows) SEP sesam backup on Windows crashes due to low disk space (VSS snapshot problem). Resolution: Will be fixed with the next version. SEP sesam v. 5.0.0.x – Proxmox VE backup does not work with the client name as plain IP address If the node added to the SEP sesam environment is not set up correctly and an IP address is used as the client name instead of the client's hostname matching the hostname returned by the Proxmox server, the backup fails. Resolution: This problem can only be solved by correctly configuring the Proxmox clients in the SEP sesam environment. For more details on configuring the client name correctly, see Proxmox VE Backup: Adding the Proxmox server (and nodes) to the SEP sesam environment. SEP sesam v. 5.0.0.x – Kopano backup containing non-ascii characters fails to create index, but backup is marked as successful A Kopano backup containing non-ascii characters results in the error UnicodeEncodeError: 'ascii' codec can't encode character... Consequently, the mail subject index cannot be created, but the backup is marked as successful. Resolution: The index problem can only be fixed by creating the correct locale: #vi /etc/locale.gen and then enable de_DE.UTF-8 UTF-8 #locale-gen SEP sesam v. 4.4.3.86–5.0.0.x – Initial Seeding functionality temporarily removed As using the Initial Seed option to set up a new Si3 deduplication store for replication resulted in an apparently successful setup, but no data was replicated, the Initial Seeding functionality is no longer available in version 5.0.0.x. It is not yet known whether it will be available again in the future.

Important fixes: SEP sesam Server 5.0.0 Jaglion

Fixed with SEP sesam Server 5.0.0 Jaglion:

SEP sesam v. ≥ 5.0.0.3-5.0.0.4 – Allowed automatic SEP sesam Client update to a higher version than SEP sesam Server Although the SEP sesam Server version should always be greater than or equal to the client/RDS version, an automatic client update is allowed before SEP sesam Server is updated. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. ≥ 5.0.0.3-5.0.0.4 – Fail to encrypt Oracle backups Oracle backups are not encrypted even though the Encryption option is enabled for the task. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. ≥ 4.4.3.84-5.0.0.4 – Exchange server backup finishes with errors and is not retried when an error such as VSS_E_PROVIDER_VETO occurs, controlled abort is not executed Exchange Server backup does not retry operation if the error VSS _E_ PROVIDER _VETO occurs and incorrectly aborts the backup, even if it is a transient error that may be caused by network errors or faulty drives that are only temporarily misbehaving. At the same time, it fails to execute the controlled abort with AbortBackup. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3-5.0.0.4 – Backup of a Windows client with the Encryption or Compression option enabled may fail with error Encrypted/compressed backups of a Windows client may fail with an error, such as MTF descriptor block ID or Unknown MTF descriptor block ID. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.4 – HPE Catalyst replication to HPE Cloud Bank store fails HPE Catalyst replication to HPE Cloud Bank store incorrectly uses sm_dedup_interface put/get instead of replicate and replication fails. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.1-5.0.0.4 – Replication for the source pool is started across multiple data stores Replication correctly selects all savesets from the source pool, but spawning across multiple data stores because the unmodified replication command does not use the correct datastore/drive where the saveset is located. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.4 – MySQL backup fails with Invalid backup source MySQL backup fails with Invalid backup source due to an incorrect check of all backup sources in the file system. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. ≥ 4.4.3.84-5.0.0.4 – Exchange server backup with VSS writer fails due to retry with 'Microsoft Exchange Replica Writer' Exchange Server backup running on the active node with VSS Writer 'Microsoft Exchange Writer' completes with an error and incorrectly attempts to run the backup again with 'Microsoft Exchange Replica Writer' and therefore fails. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3-5.0.0.4 – Backup of NFS share shows warning List of EA returned with error: errno = 95, error = Operation not supported Backup of NFS share shows the following warning for each processed file: List of EA returned with error: errno = 95, error = Operation not supported. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. ≥ 4.4.3.84-5.0.0.4 – SAP HANA backup exits with an error after a new read retry, even though everything has been backed up When a SAP HANA backup is performed and a timeout occurs, the backup is retried and all items are backed up, but the backup is still treated as unsuccessful and exits with an error. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3-5.0.0.4 – Restore from NetApp Snap Store using snapshot mount does not work Restore from a NetApp Snap Store using snapshot mount does not work due to the incorrect drive number of a data store. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3-5.0.0.4 – RHEV full CBT restore fails RHEV full CBT restore fails with Cannot add Virtual Disk. VM <VM_name> is being imported. The problem occurs because the process did not wait for the VM to be in the DOWN state. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.1-5.0.0.4 – Restore of a vSphere VM using migrated saveset fails when backup and migration were done with 4.4.3.x Restoring a vSphere VM with v. 5.0.0.x using a migrated saveset fails if backup and migration were done with 4.4.3.x. The problem is caused by SEP sesam using the wrong .sgm file. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.1-5.0.0.4 – The recorded data size after backup is incorrect After the backup is complete, the data size DB:results.data_size' is not updated correctly. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.4 – Automatic update of Windows RDS to version 5.0.0.4 may fail (Windows only) When automatically updating Windows RDS (e.g., initiated by SEP sesam Server), the update may fail and the SEP sesam service will no longer start. Workaround: Update the Windows RDS manually. If a pop-up window appears asking you to reboot the host, restart the host. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3-5.0.0.4 – Uninstalling SEP sesam RPM packages on SLES 15 fails Uninstalling SEP sesam Client and Server RPM packages on SLES 15 fails because of dependency problems. Resolution: Fixed with 5.0.0.4 SP1. SEP sesam v. 5.0.0.3 – Possible data loss due to a faulty purge function Due to the missing database entry by the purge process ('Delete everything older than the oldest Sesam day'), an FSCK error may occur as sm_pur_status could accidentally remove backup information for migrated savesets. Although the migrated version of the backup still exists, it is no longer listed in the database. Consequently, the administrator may decide to use the funtion Clean up Data Store to fix the error, which results in the deletion of the migrated saveset and loss of data. Resolution: This logic issue was addressed and fixed with 5.0.0.4. SEP sesam v. 5.0.0.3 – Possible data loss in case of migrated savesets with "lock" but expired retention Migrated savesets with expired retention period and manually set write protection in one of the previous versions are no longer write-protected ("locked") after the update to Jaglion and are immediately purged if their end of lifetime (EOL) has expired. This problem was caused by missing support for locking individual migrated savesets in version 5.0.0.3 and could lead to data loss. Resolution: Fixed with 5.0.0.4. When updating to 5.0.0.4, the lock is successfully forwarded to the backup entry. Individual locks on migrated savesets now result in a lock for all savesets of the same backup. SEP sesam v. 5.0.0.3 – SAP HANA backup is aborted by SEP sesam When running SAP HANA backup, SEP sesam aborts the backup job after about 18 minutes. This is due to the timeout setting, which is set to 540 seconds by default. SEP sesam multiplies this value and uses it as the limit in seconds for aborting running jobs (stop time). Resolution: Fixed with 5.0.0.4. SEP sesam v. 5.0.0.3 – Purging a backup on a data store does not delete metadata from the lis folder and may result in running out of disk space After a purge has finished, the lis folder still contains the metadata even though the files (.data, .info, ...) and DB entries have been deleted, which can lead to running out of disk space. Resolution: The logic is improved and the problem fixed with 5.0.0.4. SEP sesam v. 5.0.0.3 – Backup to tape with EOM fails with error Failed to detect lis file Backup to tape with EOM fails with error message Failed to detect lis file. Resolution: Fixed with 5.0.0.4. The block size is now retrieved correctly from the segment file. SEP sesam v. 5.0.0.3 – Selective restore of Windows savesets from tapes fails with "no valid MTF Stream ID was found" The selective restore of Windows savesets from tapes fails with the error message "no valid MTF Stream ID was found". The complete restore of these savesets works normally. This problem is caused by the block size not being retrieved from the segment file during a selective restore. Resolution: Fixed with 5.0.0.4. SEP sesam v. 5.0.0.3 – Citrix XenServer VM restore to selected target data store fails XenServer VM restore to selected target data store (Citrix storage repository – SR) fails as SEP sesam sbc_proxy does not use the selected SR for restore, but the default data store. Resolution: Fixed with 5.0.0.4. The selected SR is now correctly used by sbc_proxy. SEP sesam v. 5.0.0.3 – A backup of a OES NetWare resource fails After updating the OES cluster nodes to version 5.0.0.3, NSS backups are stopped due to NWSMTSConnectToTargetServiceEx CL3PROG_SERVER (NWSM_AUTH_UTF8_DATA - encoded). The problem only occurred when the source was a Linux cluster file system. Resolution: NSS requires the use of the SSL 1.0.x library. SEP sesam v. 5.0.0.3 – Exchange backup fails with error a bytes-like object is required, not 'str' After updating SEP sesam Windows Server to version 5.0.0.3, Exchange backups fail with ERROR: a bytes-like object is required, not 'str' . Resolution: The logic is improved and the problem fixed with 5.0.0.4. SEP sesam v. 5.0.0.3 – SEP sesam on Linux gets stuck at shutdown after updating (or reinstalling) to version 5.0.0.3 Jaglion After updating to version 5.0.0.3, SEP sesam on Linux gets stuck when closing the sesam processes and does not shut down, but waits for sesam service to finish for the maximum allowed time of 2 hours. The problem occurs due to the missing DB access when shutting down the system. Resolution: Start SEP sesam PostgreSQL server with 'setpriv' instead of 'su/runuser', because 'systemd' terminates 'su' sessions before stopping SEPsesam service. SEP sesam v. 5.0.0.3 – sm_shutdown instead of sm_main stop is called to stop SEP sesam on the client The systemd unit file calls sm_shutdown instead of sm_main stop on the client. Stopping the sesam service on the SEP sesam Client does not stop all SEP sesam components, especially "sm_sshd". So it can happen that not all running SEP sesam components are stopped correctly. Resolution: Now 'sm_main stop' is called successfully and the problem is fixed with 5.0.0.4. SEP sesam v. 4.4.3.75 – Tape restore performance is slower than backup performance When restoring from tape the performance is 2-3 times slower than the performance of the backup. Resolution: Fixed; fill the second buffer while processing the first buffer, EOF handling adapted. SEP sesam v. ≥ 4.4.3 – Si3S does not work in case of changed STPD port on the Client SEP Si3 source-side deduplication (Si3S) backup of a client to a SEP Si3 deduplication store not work, if the STPD service TCP port on the client side (in sm.ini and/or stpd.ini) is changed from the default port. Resolution: You can now set the STPD service TCP port on the client (client properties -> Options tab -> Listen port) to the new TCP port.

Feedback

As you come across things that you miss or are not working properly during testing, spot various bugs and issues or have another comment, please send your feedback to beta@sep.de. We will be happy to hear about your overall experience and are looking forward to your suggestions for improvement.

Thank you for your partnership and for devoting your time to SEP sesam beta testing.

See also

Release Notes 5.0.0 Jaglion – SEP sesam Release Versions