Q&A From January Webinar on Domino V12 Backup Features with Daniel Nashed

Graham Acres | 1 week ago | Comments

This blog presents the questions from the January 2022 OpenNTF webinar on the new Domino V12 Backup Features and their answers from our speaker, Daniel Nashed. Thanks for all the great questions! Again we thank Daniel for joining us for another monthly webinar.


To begin, here are the links to the HCL Domino Backup and Restore project on GitHub and the Documentation.


Q There were a couple of questions about how the new backup solution manages attachments in DAOS.
A DAOS backup is not part of the Domino backup operations. Domino backup focuses on the Domino specific part that needs a backup aware application. DAOS is a flat file backup for static files. Usually an incremental backup, which can be scheduled separately from the Domino backup.

Domino restore optionally provides a configurable interface in the same way other integrations work, to automate DAOS restore. The restore servertask detects missing NLOs and provides a list of NLOs to restore. This can be part of a restore operation and beginning with 12.0.1 there is a separate DAOS restore command for databases, folders and full servers.

Example:
load restore daos db.nsf
load restore daos mail/
load restore daos -all
Q What about the files placed in DAOS - then you restore some deleted documents ?
A In a best practice configuration the retention time for the NLO prune is longer than the backup retention time. By design NLOs stay on disk for that time and will only increase the ref count for the NLO again, when the document is restored.

In case the NLO is not present, there are integration options for DAOS restore automation. See the previous question.
Q Does this full backup, backup NSF's and NTF's? Anything else?
A The backup is designed to backup all types of *.nsf and *.ntf +*.box files only. Beside that in archive transaction log mode, TXN files are separately backed up like for any other Domino integrated backup solutions.

All other files in the data directory have to be backed up by a file level backup -- this includes notes.ini and server.id. In case of a snapshot backup, usually all files are in backup. But Domino restore does only provide a way to restore all type of NSF files.
Q Can the restore function fix a folder where someone has removed all the documents from the folder but didn't delete the documents.
A We tested this to confirm. The answer is No. Restore is intended to bring the folder back along with documents, but not to resynch folders if they are not deleted. The "Delete" part of this is the important trigger.
Q What if we need to restore some design elements? Which are documents also. Is it possible? Some kind of app version control system?
A Design elements are part of the databases and are in the backup. There is no special restore mechanism for design. And there is no restore version control system.

Design elements are "notes" as well. Not documents. They have a different note class and are handled in a completely different way in Domino than documents. This includes a different way to replicate them. Special restore options for design elements are not feasible to be implemented. There are too many edge cases. And the restore operation is optimized for the main use cases: restoring end-user data including documents an folders.

You can restore databases with changed replica ID or disabled replication and use the new Domino compare application to compare the design. But you have to change the design manually. There is no out of the box functionality.
Q Can the restore function - restore just the design of a database. I'm thinking about application updates that go wrong that the database can be brought back to a previous design.
A No support for database design restores. However you could restore a database, make it a template and refresh the design of your original database.
Q I may have misheard during the presentation: When restoring from backup, a doc is given a new UNID?
A Documents are restored with the same UNID and the UNID is used to identify the deleted document form the deletion stub. Only the sequence time and sequence number is updated to ensure the document will be replicated to another database instance.
Q What is the recommended way/config for 5-min-deep backup? What if my (crazy) boss wants to loose docs for not more than 5 minutes?
A Running a backup every 5 minutes even for a single VIP database is not needed. You can point in time recover databases up to the minute leveraging transaction logging.

1. Restore last database backup
2. Roll forward all transactions logs created between the full backup and the point in time.

Note: Point in time with a long delta time can cause extensive restore operations and will result in all TXN files to be examined for transactions for the requested restore database. Best practices is to have multiple full backups per week. With deduplicating storage a nightly backup will cause minimal storage overhead.
Q When does/would a DBIID change?
A 1. Fixup -j
2. Compact -b/-B/-D, DBMT compact
3. A restore operation changes the DBIID as well if the original database is still on disk
Q Is it possible to restore prod backup to another server just to test that everything in production is set up correctly and works properly in high-loaded production environment?
A Databases should be always restored to the same server. There are too many edge cases where a restore has to be on the same server (DAOS, Translog, etc.). An updated restore configuration would be possible, but is not easy to configure nor it is supported.

The use case described is also not a good test. A restore operation without translog roll forward is a light operation for the server.
Q If I have a secondary backup system for GFS retention policy, what I've to do to restore a database already purged (including NLO and archived translog) from the directory store of domino backup?
A There is no direct integration. You can restore NLOs if they are still present on the backup environment. A restore always needs the inventory for a backup to have a pointer to restore the database. The meta data gets purged when the backup is purged for backup restore consistency reasons.

Disabling the purge can be configured, but is not recommended. You can restore the backup inventory from a log file in case you want to keep the databases for a longer time. But there is no integration that allows an automated restore nor that it is a supported environment.
Q How does the 'exclude databases (pattern)' option work? How to specify only one or 2 key applications to back up?
A We will post export and import examples shortly on GitHub. It can be done but the exclude patterns are designed more for flexible exclusion. The exclusion can be inverted.
Q Can I have a backup configuration for just a couple of directories on the server? I would like to just backup the mail directory and not the entire server.
A See previous question. This would be using more complex exclude pattern.

If you want a selective backup for one time only, you can just specify a single directory or database. This will always include the sub-directories of the specified directory.

load backup mail/
Q What happens if dominobackup.nsf becomes corrupted?
A The database is always in backup. But in theory it can also be corrupted. There are two parts which are important in the database.

1. Backup configuration

The configuration is exported to a DXL file every time it changes. And can be added to the backup to be restored into a new dominobackup.nsf

2. Database inventory documents

The database inventory documents for a backup can be rebuild from the backup log as well.

You can specify a different database for a restore in this case to not conflict with an existing database.
Q When I restore a file with a new replica ID and I have a replica job between 2 server get I than a replica issue with this DB?
A The restore operation by default disables replication and you would need it enabled in the restore document when you want to replicate the database. If you decide to change the replica ID, this is the safest way to never let the database replicate with any other database in your environment. In either way, it will not conflict with replication.

Answers by Daniel Nashed. Edited by Graham Acres.



January OpenNTF Webinar - Backup your Domino Server - New Options in V12

Graham Acres | 1 month ago | Comments
Domino 12 introduced a new and very flexible Backup solution to bridge the gap between Domino and backup applications.

This session provides a jumpstart into this new functionality and technical background to understand the different types of integration options. Learn about the new backup feature in Domino 12 and discover how to integrate widely used backup solutions like Veeam. Watch the new backup feature in use with a live demo.

This will be a great session if you haven't been backing up your Domino server or are already using other backup solutions and want to integrate them better with Domino.

Your presenter will be Daniel Nashed from Nash!Com. He will answer your questions at the end.

This webinar will be held on Thursday, January 20, 2022 at 11:00 AM to 12:30 (New York Time).

Register Here

OpenNTF 20th Anniversary Celebration

Graham Acres | 1 month ago | Comments
We hosted a celebration of the 20th Anniversary of OpenNTF on Tuesday, December 14, 2021. Thank you to everyone who joined. It was great to see everyone and share some good stories.

Special thanks to our guest speakers Bruce Elgort, Niklas Heidloff, Justin Hill, Christian Gudemann, Devin Olson, Paul Withers, Ulrich Krause, Vince Schuurman, Jesse Gallagher, Julian Robichaux and Oliver Busse for sharing their stories about OpenNTF and some of the projects that have helped the community over the years. Thanks also go to Graham Acres, Serdar Basegmez, Howard Greenberg and Roberto Boccadoro in addition to the other Board members who spoke for helping to make this celebration a fun trip down memory lane.

Links are included at the bottom of this post to the special video tribute shared at the start of the celebration, prepared by Serdar Basegmez, and to the actual celebration video itself. We were scheduled for 90 minutes but ended up staying on for 2 hours telling stories.

This really is a wonderful community, and we thank everyone for taking part and sharing. Code. Learn. *Break* Share.

Anniversary Participants Screenshot

Link to YouTube video of the 20th Anniversary Webinar Link to YouTube Video of OpenNTF History