Challenge 4: Business Continuity and Disaster Recovery with Azure NetApp Files - Coach’s Guide

< Previous Challenge - Home - Next Challenge >

Notes & Guidance

Section Sub-Section Comments
Backup Adjust Log Backups volume size and HANA volume path change Change it to 250 GiB from the initial 100 GiB size. This will match with daily log backups requirement. Use HANA Studio to change the log backup basepath parameters - change for both SYSTEM and Tenant DBs. Parameters Volume Size
Backup Log Backup Timeout Same method for changing as the log backup volume location parameters. LogBackupTimeout
Backup Build snapshot solution (azacsnap) Using azacsnap. Install azacsnap from the repo provided. Install the tool on the management VM/Jump Server Linux, update config JSON, and add to cron tab. You only need one config JSON for this. Put data volumes in the data section, and log backup in “other” section of the config JSON. Refer to the ANF Blogs (no need to use aztools referenced in the blog. Azacsnap for data would run twice a day (pick your time, say 6am and 6pm). Log backup frequency of 15 min (900) will yield log backups every 15 mi, but schedule (in cron) it a minute past the log backups i.e. 16th, 31st, 46th and 01st min of the hour. Common pitfalls are missing the hdbuserstore key step or entering incorrect hdbuserstore key name when the acacsnap tool asks for it and not placing the azureauth.json in the $HOME/bin directory.azacsnapConfig azacsnapCron
Backup Execute adhoc snapshot Since the crontab schedule may not immediately trigger snapshots, we want to take an ad-hoc data volume snapshot so that we can proceed with the next steps of offloading to blob container. Use the azansnap command you wrapped in the crontab and execute it thought the command line. azacsnapAdhoc azacsnapAdhocSt azacsnapAdhocPo
Backup Offload with azcopy Create a Managed ID for HANA VM and assign Blob Owner, Reader and Contributor permissions for the Blob Storage Account. Install azcopy on the HANA VM. Run once to show how the on-demand azcopy would move both data and log backups over to respective containers. For data volumes, you will sync the contents of the .snapshot directory under /hana/data/SID/mnt00001/.snapshot, and for the log backups, you will sync the actual log backups files for SYSTEM and tenants by simply syncing the entire /backup/log directory. Refer to the blog for more information. SLES 12 has a bug and may prevent you from running azcopy in a managed id session. You will need to run azcopy login --identity, followed by keyctl show, you will find no azcopy key at the moment. Run keyctl session workaroundSession, then run this keyctl show again, and now you should see the azcopy related key. Now you are good to proceed with azcopy sync command. azcopyLin azcopyPortal
Backup Retention for blobs Add a lifecyle management policy rule to delete the blobs filtered for the two containers to be deleted if modified older than 7 days. Blob Retention
Backup Restore test This is a file level recovery test. When reverting, choose revert to volume ANF option so you dont have to update the fstab.
DR Assess Requirements No action, just laying out requirements. The DR region is East, so that’s where the ANF replication will be outbound for. Since, RPO is <30 min, we will use 10 min CRR interval.
DR Set up CRR Refer to the blog and CRR MS Docs reference link. Set up replication for data and log backups (we will not be needing to use shared volume for this OpenHack). Set the frequency for replication to be 10 min. CRR Replication
DR Create a placeholder file We will not be validating the availability of this file or the security user as part of the validation. This is just in case the participant decides to do the optional steps of provisioning a VM, installing the HANA DB, and then doing the recovery at the DR site
DR Take an ad-hoc backup Use azacsnap
DR Execute DR Without deleting the replication, you will not be able to revert to a snapshot on the existing replicating volumes. When changing performance tiers, create a new storage pool of premium tier first and then move the volume from standard to this new premium storage pool. This will conclude the DR validation from the challenge’s perspective

Learning Resources