Backup SQL Database with Cohesity DataPlatform: Part 1

hi my name is Abhisit chat anchor and I'm an engineer at cohesive and today I'm going to talk about the co ICD data platform in particular the sequel integration that we have so as you know Kohi city is a distributed hyper-converged secondary storage platform for all your secondary workloads it is driven by the snap tree technology which enables backups that are fully hydrated yet space efficient the platform the data platform also provide also provides global deduplication compression and encryption for all your data Kohi CD also provides data protection capabilities which include backing up your VM workloads physical servers and database database protection such as Microsoft sequel several database and today we are gonna talk about the sequel protection in in more detail so sequel backups come in two different flavors the first one is sequel Falls and the second one is sequel log backups sequel full backups they make use of the VSS API and they trigger VSS snapshot using the VSS API once that snapshot is taken we use the change block tracking mechanism or CBT to incremental e backup the snapshot and so why do we do that because it keeps the backups fully hydrated on cohesive and it also eliminates the need for doing sequel differential backups sequel log backups are needed to backup transaction logs and in this case we use the VDI interface to do to backup the transaction logs and these sequel long backups can be used for doing point in time restores now if you're going to talk about sequel backups in a virtualized environment so our sequel backups in virtualized environments are agentless and we use a helper app we actually dynamically load and unload a helper app to orchestrate the backup flow so suppose that you have a sequel server host right here which which is running sequel database which we you want to protect you would also be running a VSS service and a VDI service which would be interacting with your sequel and now we want to protect this sequel server so the first step in doing so would be you need to register the sequel server so this involves specifying credentials which might be used to in to unload and load our helper app on the sequel server host once we have the right credentials which we will we will load the cohesive helper app which will then interact with the VSS service the VDI service and with the sequel server directly to fetch metadata and I'm going to talk about that now in order to load and unload the sequel server and to perform some VM specific operations we make use of V a DP interface that's provided by the VM there and once the helper app is up and running on the sequel server we would also interact with that so the helper app acts as a dump agent and all the orchestration is being driven by the cohesive cluster so let's go through what are the steps that are involved in a full backup so for a sequel full the steps that we would take are of course we would first load the helper app once the app is up and running we would take a VSS snap snapshot and this involves interfacing with the VSS service that's already running on the windows host once the VESA snapshot is taken we will trigger a VM snapshot this is the incremental snapshot and VMware can now give us the change blocks which we would incremental backup so the backup is fully hydrated on Chloe CD and yet it is space efficient once the VM snapshot is taken we'd actually back it up as I just described once the snapshot is backed up we would actually fetch the metadata that's associated with the visa snapshot this metadata includes things like Alison or lock sequence numbers globally unique IDs associated with the database and so on and this information would later be used to perform a DSS restore once the metadata is saved on ko he city we would then go ahead and release the VSS and VM snapshots and in the end we would unload the agent and that's your secret full backup the sequel log backup flow is fairly similar to that of full as before first we load the helper app in this case we interface with the VDI to perform the transaction log backup once the transaction logs are backed up we fetch the metadata and save it on Co history and this metadata includes lock sequence numbers goods and so on and so forth once the metadata is saved we would unload the agent and that's it once the sequel falls in log backups are on Co a CD platform they enjoy the benefits of the platform for example the full backups are fully hydrated the data is globally duped compressed and encrypted that together with the metadata that we have saved allows us to support interesting restore scenarios like point in time database restores restoring the full DM and also various tests under workflows but that would be a topic of a different video you

Channel: Cohesity

Views: 4,774

Rating: 5 out of 5

Keywords: SQL backups, Virtual Device Interface (VDI), Volume Shadow Copy Service (VSS), Change Block Tracking (CBT), full backups, point-in-time restore, transaction log backups, development and testing (DevTest), fully hydrated

Id: f33U3tZH1rE

Length: 8min 15sec (495 seconds)

Published: Mon Jan 09 2017