Hello and welcome to Dell Technologies Power Protect A P series appliance. Our my Data domain capacity resolution path. In this video, we will talk about space reclamation on data domain referenced article 192662 section space reclamation.
This is applicable when you intend to manage the data domain capacity are when you high utilization is seen on data domain on when a maintenance starts failing due to data domain being full and a starts receiving integrity alerts due to the same. This video was created to get familiar with digi domain cleaning.
Like when to go for space reclamation, how to perform DD cleaning, how to schedule DD cleaning for it to run automatically. Details regarding DD cleaning, throttle and frequency, how to check avo maintenance status, how to verify checkpoints on AVO and snapshots on DD.
Now let's quickly recap on what we learned in part one and then see details regarding space reclamation along with demo, what we have seen so far is uh we need to first check the utilization team device by logging into the data domain via AC M or directly and then check uh which tier is having high utilization.
And if the particular tier is having a high utilization, then is there a cleanable size present for that um Tier or not? If there is, then uh we will go with the space declamation. Let's see what space reclamation talks about. So space reclamation basically means that the DD cleaning process is the one that is responsible for reclaiming space on data domain against the expired backups. OK.
So if active tier utilization is high, we will initiate cleaning on the active tier. And if the cloud utilization is high, we will initiate the cleaning on the cloud cloud tier. And how you're going to do that is either you can use the command file as clean start for active tier and cloud clean start cloud unit name for the cloud tier or you can also do the same through the U I.
So let us see how you can do the same through U I. The first thing you need to know here is the active tier cleaning and cloud cleaning. There are certain default schedules that are present on data domain. So you're going to check if you know, customer has reverted those schedules to never.
Is it uh actually indeed being present the schedules is the are the cleaning cycles being run regularly or not? So that can be checked once you log into the DDU I again through AC M or directly, you will go to the same data management file system tabs from the left panel and then go to the settings gear I can see at the right side. Ok, click on it and then go to the cleaning tab.
Here you will see regard active tier cleaning and cloud tier cleaning. So the schedules are actually can be set over here can be, you know, checked from here whether it's present or not. Uh, here we can see it is weekly. The frequency is for active tier cleaning is weekly, 6 a.m. Tuesday, which is the default schedule.
And for cloud tier, the default schedule is actually for after every four active tier cleaning. So cloud cleaning is dependent on active care cleaning after every n number of active care cleaning, uh, a cloudier clean can run. Ok. So either, uh, you can change it uh, to two or, uh, let it be as default four, but it needs to be present so that the cleaning cycles are run automatically and there's no, uh, require, uh, there's no need to do manual intervention anytime, ok.
As well as you see something called throttle percentage over here. This is the amount of CPU cycles that you would like to assign it to the cleaning process. 50% is the default throttle, but in instances where you would like to increase it for a certain amount of time, like for one particular cycle, like in this case, your, you know, a maintenance has failed, the PD is full, for example, it's 99% full or, uh, you know, and you want the, and you do see cleanable space, a lot of cleanable space, that means there are expired backups.
There is scope of reclaiming space. Uh And you want to give priority to the cleaning process so that it runs with it, you know, in lesser duration than usual, um, and reclaim space quickly. Ok? And in that case, what you do is you put your backups on hold and because there's no space on the d any which way to accommodate any more backups.
Uh So you put your backups on hold and increase this sort for, let's say 80% ok? And you wait for the cleaning cycle to complete. Remember whenever the cleaning cycle completes, you got to reward back the throttle back to what it was earlier, either you can keep it to default.
Uh, but don't, uh, you know, leave it to the increased throttle because that can impact the performances for the other process that would be running uh on data domain. Ok? So if you want uh every process to run parallel then do not, uh keep this throttle, uh you know, uh at a higher level.
And also the frequency that you see over here that from our perspective and from product perspective, basically, we uh suggest not to set it too frequently. Once in a week is default. Um, for active care cleaning, once in a week is default. But if you want to, you know, increase the frequency.
See, then you can do it maximum twice a week, that should be fine. Uh but do not go for too frequent cleaning cycles because that would impact the data locality present on data domain. OK. And that uh it causes high fragmentation and that would lead to um impact on uh the rep performance as well. OK.
That is the reason you do not have the cleaning set too frequent. Uh, that was all about cleaning schedule, uh, throttle as well as its frequency. Ok, coming to, uh, back to our point where your, you know, organization is high and your maintenance is failing. You would like to start the, though the cleaning schedules are present, you can't wait for the cleaning schedule to, you know, until the time the next schedule run.
Uh, what you would prefer in that circumstance as you would want to initiate the cleaning cycle immediately and how you can do that from U I. You would see an arrow over here at the bottom at the right side. Ok, expand that here. You will have an option for active care cleaning and a start button against it.
So, clicking on this, the cleaning will get started immediately and if you want to alter the throttle, then please please do that before initiating the cleaning cycle. And once cleaning cycle completes, you can revert back the throttle, uh, from the same place from the settings gear icon. Ok.
The same thing can be done from the CLI as well. So if you want to see the, uh, schedules, you will see clean show schedule. Ok. You will see for cloud cleaning, you will see cloud clean show frequency, uh, sorry, frequency show. Ok. And then, uh, same goes with the throttle files as clean, uh, show throttle and cloud clean throttle show. Ok.
And if you want to set it reset it, you may instead of show command, you may uh you set command to set it to the value required value that you want to run for. Uh So these are the schedule throttle. And then if you want to initiate the cleaning cycle, fal clean start as the command for activity, cleaning cycle and cloud clean uh start cloud unit name.
So let's say for example, cloud unit name is cloud one. That would be your cloud clean start command to know what cloud units are present to even run this command. Like for example, uh I have taken cloud unit name as cloud one here. So what are the cloud unit names on your device? You can check it through the command cloud unit list uh to get the cloud unit name and then run the cloud unit.
Uh cloud clean start command uh with the cloud unit name. Ok. And then to check uh where you know cleaning status you can do for active tier cleaning, you can do fast clean status. And if if you want to see if the cleaning is uh you know, ongoing. If you want to see the real time uh progress of the cleaning, you can do the watch command.
This is for the active tier, same goes for the cloud tier cleaning as well. Cloud clean status and cloud clean watch for the real-time uh progress. And to come out of the watch commands, you can do control C OK? So this was all about uh the cl A commands that would be helpful and all of them are documented in the KB articles that are referenced over here as well as the KB also talks about uh the various faces.
The cleaning goes through what it does uh the activity of cleaning. OK? Um One thing to note here is active tier cleaning as well as and this cloud cleaning are mutually exclusive process. Only one can run at a time. They cannot run in parallel. OK? If let's, you know, continue with the same exam as we had started.
So your activity utilization is high and your requirement and preference now is to start active peer cleaning. But when you check on the system, even if you try to started, you will get an error that cloud cleaning is in progress. So you can check the cloud clean status to see what phase it is on. Um And if you would have, if you have preference here, for example, active care utilization is high, you would want to start the cloud cleaning.
But even though you want to start, you are unable to do that since cloud cleaning is in progress. So what you would do in that scenario is you will go ahead and stop the ongoing cloud cleaning. So how you will do that is cloud clean stop? Ok. Use this command and verify using cloud clean status.
Wait, uh wait for cloud cleaning to stop and reflect in the output of cloud clean status. Once that is done, then you can go ahead with false clean start. That is your active data. OK? So always check the status before you start any tier cleaning as well as remember both the tier, uh both the tier cleaning cannot run in.
So uh all this is uh documented as well in this uh reference article as I was saying, you. So now once the cleaning is initiated, you need to wait for it to complete and then check whether it has been successfully completed. F 3D cleaning has not successfully completed.
You can contact support and we can guide you further. Otherwise, once cleaning gets successfully completed, check whether the space is ok, post cleaning or not. OK. And how to check space? We have already seen that in the earlier slide, you check through DF uh our files shows space. OK? Now, once that is done, you see the space is OK or not post cleaning.
If space is OK, post cleaning, then you got to wait for the next successful a maintenance cycle. Ok. And then re run DD cleaning. And why is that? So, because, uh, earlier on the maintenance cycle was failing. So if there were any backups that are expired, uh, manually or if they have met retention, uh, then that has, uh, that has not been communicated to DD yet.
DD doesn't have any information since the time the earlier maintenance was failing. If there were any backups that have been expired DD has no information about it, it only receive information, uh, once a, uh, a maintenance is done. So, uh, you need to wait for the successful a maintenance which is scheduled to run every day in the morning on a, uh, and that includes the AAR checkpoint HFS check and GC. Ok.
Once that is done, uh, you can, uh, rerun the DD cleaning using the same commands that we have seen earlier. Ok. And how you can check about this maintenance cycle, whether it has been successful or not, you can log in to the cli, ok. And run the command called status dot DPN. Ok.
So this will show you the information about all of those three processes that are included in a maintenance that is the checkpoint gcnhfs check. Ok. All of them states. Ok. That means the maintenance has run successfully. Ok. And you can also see the CP list command to see the checkpoints that have been taken. Ok.
So uh A contains maximum uh two checkpoints over here and that should be the latest from the uh you know, maintenance cycle that has been run today. Ok. So G 12. So all this, this checkpoint and the rolling checkpoint has been created uh when the Aboma maintenance ran today. Ok.
So you got to check if the status is. Ok. And you have the checkpoints from today. Once you confirm your Aboma maintenance cycle has run successfully, you can go back to DD. Uh and then um you know, run rerun the DD cleaning as we have seen earlier. The another verification that you can do do before initiating the cleaning is you can log into the uh data domain cli uh see the entry list, fetch the AAR entry name and then see the snapshot against this uh entry. OK?
And this should match with the uh checkpoints that you are seeing on the AAR. So whatever active checkpoints, active snapshots are seen on DD, those should match with the uh checkpoints on AVO rest all should show as expired. So that means a successfully communicated to data domain through today's maintenance.
So that part of communication has been completed and now you are good to proceed with the data clean. OK. So, and you do that through fastest clean start or cloud clean start unit, cloud unit name as we have seen earlier.
So again, you would wait for the cleaning to complete and check the space is OK? Post cleaning or not using the same command files that show space or DF on data domain. OK? And then once you see that it's all fine now it has reclaimed space uh from whatever has been expired so far.
Uh As the AVO maintenance has completed successfully and communicated to data domain. So it has taken full advantage of the cleaning this time and it has reclaimed everything it can. OK? So that means now it is time to acknowledge the alerts on a since the maintenance has done successfully.
Um And the maintenance ran successfully because the earlier DD cleaning was also successful and it proclaimed amount of space for the maintenance to go through successfully.
And so we are now going to acknowledge this alert and to acknowledge this alert, you uh will have to, you know, contact support and we can do that for you and that will resolve your capacity concerns.
