Storage and Deduplication

Hi,How may I know if my Aix environment is supported with DataDomain DDBoost over fiber?  Commvault Version: 11.40 release 22Commvault Media Agent SO: AIX 7.1/7.2 We have a large Oracle DB with 50Tb that we would like to protect.

Hi all, I’m planning on replacing our MA server, and I just needed a sanity check on the Replace Media Agent App and Move Mount Point processes. The old MA has a Disk Library with 7 mounth paths, most of which are iSCSI LUNs from a couple different arrays. The new MA will have much more direct attached storage, so I’m plannin on just 2 mount points. If I run the Replace Meda Agent App process, and let it run the Move Mount Point process, am I able to consolidate the old mount paths down to 2 during this process, and specify the destination of each of the source MP’s?  I assumed that I can because the Move Mount Path function in the Commcell console offers me the option to move it to an existing device, however the documentation says: The move mount path operation overwrites the destination mount path. Therefore, ensure that the destination mount path is an empty folder.  Can anyone offer any clafification?Thanks!

Can Commvault store full backups and incremental backups in separate buckets, with immutability enabled only for the full backup bucket, to optimize costs? If feasible, what are the pros, cons, and key considerations of this design?

The topic generally covers a few points.We have an “older” backup system, running Windows Server with “suitable” hardware specs, containing >50 TB of storage. No cloud stuff etc. but an LTO-9 library attached.Why is the fragmentation so high at all. I’ve configured allocation of writeblocks to 1024 MB. But it seems to clearly not do - or keep - that. I hoped to find help in:  But the recommendation to use “space reclamaition” instead of defragmenting sounds like someone has no clue about the topic. If my library is “full”, I can’t reclaim anything and by that it does not help the fragmentation-status the tiniest bit. For DR-use, we not only save the dr-backups (config etc.) but also have a scheduled job running, that backs up (copies) all library-files from storage to our tape library (also enabling us to airgap data frequently - good for ransomware attacks). But due to the fragmentation, our backup process is limited by the RAID containing the library. It’s still “spinning rust”

Hello community, we used a deduplicated WORM cloud storage pool  for a while but needed to start fresh because of unexpected growth.In the process of aging the data of the old pool we noticed an issue because of the default 1 cycle retention which required a workaround to bypass the cycle retention.Therefore we would like to enable zero cycle retention on the new WORM pool like described here: Configuring WORM Storage LockWhen preparing the pool and assigning the first storage policy to it I noticed that the cycle retention got increased to 1 after enabling WORM despite running the command from the documentation beforehand.So my question is: Is there more to do than described in the documentation to enable the zero cycle retention on WORM pools or is the GUI just “showing” 1 cycle but using 0 in the background?

What fields/tables  are used to decide the copy type of an StoragePlicyCopy : “snap primary”,”primary”, “selective”,”synchronous”,… 

We have a client sitting in a DMZ that can successfully connect to media agent 01. However, media agent 01 does not have a DDB. In the job logs we can see the scan phase for the job complete on MA01 but then it fails on the backup phase. We noticed in the logs an attempt to connect to the CVD service on media agent 02 which houses the DDB. The client does not have a path to MA02. In this situation, does commvault require the client to have comms to the DDB media agent MA02? I thought that MA01 would handle connecting to MA02 not the client. How does Commvault handle the data flow in this setup?  Thanks!

Hi all, I would like to enable client-side deduplication on a Windows VSA proxy so that only unique data is sent directly from the VSA to the MediaAgent DDB/backup library, instead of sending all read data from the VSA and deduplicating only on the MediaAgent side.We are looking to this approach because the VSA proxy will be a VM inside an AVS environment, while the MediaAgent will be outside of AVS (as Azure VM).  Could you please help with:Required packages on Windows VSA proxy: What Commvault packages are needed on the Windows server acting as VSA proxy to support client-side deduplication (e.g., Virtual Server, File System Core, File System, or any additional components)?How to enable client-side deduplication for VSA backups: What are the step-by-step configuration steps (storage policy, subclient properties, etc.) to enable source-side/client dedup specifically for VSA workloads?How to verify client-side dedup is working: What are the step-by-step checks (job details fields, repo

We have decided to move our backup data to the NetApp AFF A90 and allow the existing NetApp system to age off in line with the retention policy.Could you please advise on how best to present the volumes to the MediaAgents—whether as block storage or via NFS? We would like to understand which option would be more suitable for our use case.Additionally, could you confirm whether NVMe is supported by Commvault version 11.40?

Dear Team, We recently migrated our CommCell server using the "Recovering the MongoDB Database Using the Mongo Recovery Assistant Tool for Windows" document. After the migration, the new CommCell server (Commsafe) cloned all configurations from the old server, including the old Disk Library, Mount Paths, and DDB entries.We have since configured a new Disk Library, new Data Path, and new DDB on the Commsafe server and successfully re-associated all subclients to the new Storage Policy. The old CommCell server has been fully decommissioned, and the old client win-i3vamh7fvl3 has been retired from CommCell Console.However, when we attempt to delete the old Storage Policy, we receive an error stating that DDB Backups are still associated with it. Additionally, when attempting deletion, the Job Controller shows pending/queued jobs tied to the old Storage Policy, which is blocking the deletion.What we need help with:How to clear the stale/orphaned DDB association left from the cloned configu

Hi,according to BOL (ReFS for Windows 2012 and Windows 2016), ReFS is supported for Disklibraries up to Windows Version 2019.Is there a reason to not extent the support to the recent versions of Windows OS ? (2022, 2025)Is it required to copy-out, reformat using NTFS, copy-back all disklib paths/volumes, if the server is upgraded from Windows 2019 to a more recent version of Windows ?This can become challenging if the disklib has not eough freespace available to temporary evacuate one ot its paths.BRKlaus 

Hi community, we recently deleted a deduplicated storage pool which was using a S3 bucket on our NetApp Storage Grid and had WORM/object lock enabled.This was nearly a week ago and we still see almost no space beeing cleared on the bucket.We are seeing delete markers being set (sometimes even multiple for some reason) but older versions with actual data are not being deleted.Also Commvault is still issuing S3 delete commands to the storage and the delete marker count is already higher than the count of actual object versions which seems weird. We noticed a recommendation to set up a lifecycle policy on the storage to delete non-current-versions which we did and now is clearing up some more space than before but still very slowly, but this is more of a storage side issue.Source: Configuring WORM Storage Lock We noticed correct deletion behaviour (physical pruning) in the past on other active storage pools with S3 object lock so my question is:Is it necessary to set up these lifecycle po

I came to a client recently that had a large backlog of aux copy data and problematic tape libraries. After some remediation I got the tape copy running on 3 drives while awaiting replacements for a couple of broken ones. The replacements were done and following the resumption of the backup the tape copy job will only select one drive. So there are 2 media agents, each with a tape library attached, one has 3 drives and the other has 2 drives, so 5 in total. The storage policy has the 2 libraries in the path and 5 streams with multiplexing x 25. The drive pools on both libraries are set to use all drives. I cannot understand why it will only use a single drive since the backup job was resumed where it was using 3 previously. It now has 5 working drives which can all be seen in Commvault and all work fine in the Tape library GUI’s tests.Commvault version is v11.32.115. 

Hello everyone,I have a question.Since we don't want to renew our maintenance contract for our on-premises backup storage, we're now wondering whether we can run the entire backup data set, which is 89 TB, on AWS S3. Has anyone had any experience with this? What about read accesses during data verification and data aging? What kind of costs should we expect? Commvault mentions a value of 20 TB for data verification jobs, for example. However, I don't think that's realistic, since the data verification job only reads blocks. To reduce read costs from S3, there's supposedly the option of deploying a media agent in AWS to eliminate these costs. Does anyone have any experience with this?I'd like to get a rough idea of ​​the ongoing costs we can expect if we were to go this route.I would really appreciate any feedback.Regards, Thomas

Hi, We have a full Linux Commcell (Redhat 9.5) in V 11.36 with a 2-node Grid MA and partitionned DDB.Firewall is enabled and port 2049,2050 and 111 are open.When I start a ddb Verif, it goes to 50% successfuly, (local verfification) and goes in pending when it tries to validate the network path on the other node using NFS.If I stop the FW on the MA then the ddb verif finish with success :BEFORE :20679 51f2 04/23 17:43:04 #### initiateSession: 3dfs server @ [MA002] started successfully!!20679 51f2 04/23 17:43:04 #### initiateSession: Share MA002:/ma/MA002/vol01/JKOLPE_04.23.2025_11.14/CV_MAGNETIC exported successfully as /ma/MA002/vol01/JKOLPE_04.23.2025_11.14/CV_MAGNETIC_c1e0710d-b14e-46aa-b840-2e833e4a4d60_21 for session [22]20679 51f2 04/23 17:43:05 #### waitForNfsReply: Error: nfs_service failed for Session [22], revents: 0x1C err: -1->rpc_service: socket error No route to host(113).20679 51f2 04/23 17:43:05 #### didMountSucceed: Failed to mount MA002:/ma/MA002/vol01/JKOLPE_04.23

Hey all,We had two media agents on site A and site B. Site A’s media agent died and was replaced by the media agent from site B. Site B was consolidated. I have a Storage Policy GDP_xx that points to the dead media agent.I tried to replace the  MediaAgent in the Data Paths of the Properties windows, on the primary copy of the policy (the only copy, there is no other copy) We have a Dedup DB reconstruction job for this policy GDP_xx.How to stop this auto job from running? How can I disable the Storage policy?

We’re currently in the process of doing a major upgrade of our Commvault setup - both software version and hardware, and are looking at ways to improve things. None of us are seasoned Commvault admins - I did a course before starting a couple of years ago, only to find that much was irrelevant as the course focused on the webGUI and Plans, while the existing setup turned out to be “old style”. I’ve also had a lot of other stuff on my plate, so there has been little time to do much self study...So we’re taking this opportunity to build the setup from scratch. One of the areas we want to improve is protection against malicious or accidental deletion. The old setup does create aux copies to an offsite disk library, but as both primary and secondary libraries are permanently mounted via iSCSI it would be trivial to obliterate everything… an “interesting” fact here is that the previous incarnation of the setup used NetApp for storage, and SnapMirror to create offsite replicas - so in some w

Hello Team, Please advise the best practice for the implementation of on-premises HPE Object storage with Commvault.There are two different sites, one site will use HPE Object storage for backup and another site will use HPE File System for backup targets (CIFS shares mount paths).Our requirement that backing of VM’s, FS & DB’s, the backup data of around 500TB.Is it single (or) multiple S3 objects will create from Object HPE storage. There are two physical media agents to use these object devices access. Shall we create single Disk library and will add mount paths S3 objects. Any best practices to configure additionally to get backup performances. The backup data will copy from Object to File System storage using Commvault Aux copies. Is it will work for these copy jobs.Thanks in advance.

Dear community,I have an issue after migration from Windows 11.36.49 to Linux 11.36.49 , the Vault Tracker reports is currently empty :No tape from the location appear in the “Due backup for reuse” report while 7 tapes get a status “Now” on the “Due back for reuse on” field from list of media : On the vault tracking report with broader criteria, I get the following message : Please, what do I have to checkBest regards,Gilles

Good morning,We are experiencing an issue with our Commvault environment related to copying data to tape, and I am hoping someone can help explain the behavior or suggest a fix.Currently, our Commvault system is configured to take weekly FULL backups of all file shares and VMs every Thursday. Once these full backups complete, they automatically copy to tape. This process itself works as expected.The problem is the volume of data being written to tape.Each week, I have to manually update the Copy Policy “From Date” to the Monday of the current week to ensure that only the most recent data is copied. If I forget to do this, the amount of data copied to tape increases significantly.For example:January 2026 Week 1: I set the copy date to Monday 5th January. On Thursday, approximately 100 TB is written to tape (usually completing by the following Wednesday). Week 2: I forget to update the date. On Thursday, it again attempts to copy around 100 TB. Week 3: I forget to update the date aga

Incremental Backup Converted to Full (7.5TB+) following Synthetic Full Completion Problem DescriptionA recurring issue has been identified where incremental backup jobs significantly increase in size (bloating from ~1.5GB to 7.5TB+) specifically after a Synthetic Full backup is performed. In the most recent instance, the backup remained in a "Running" status for over 10 days. While the job continues to progress slowly due to limited bandwidth between the Cloud source and the on-premises Media Agent, the root cause is a logic failure in the backup chain.Technical Analysis Automatic Job Conversion: Commvault automatically converts an incremental backup to a full backup if it cannot establish a valid baseline or if the previous cycle did not close correctly.Cycle Disruption: A backup cycle starts with a Full/Synthetic Full and ends before the next Full. If the Synthetic Full (DASH Full) fails to properly sync its metadata index or deduplication signatures due to network latency, the subse

The M365 admin center reports my OneDrive as having 2.8M files consuming 8.1TB of space.But when I go into Commvault I see numbers that are totally different. In Commvault I see either 1) the latest backup had 3.2M active and 0.7M Inactive for a total of 3.9M files or 2) all backups with 6.6M Active and 2.0M inactive for a total of 8.6M files.Likewise, the disk space used is reported as either 1) the latest backup with 3.4TB Active and 1.6TB Inactive for a total of 5.0TB or 2) 9.6TB Active and 4.0TB inactive for a total of 13.6TB.Is there any way to drill down into Commvault to figure out why the numbers – particularly the “latest backup” file counts – are so different from what is reported by Microsoft?

I want to configure a disk library and share it between two Media Agents running Windows Server 2022, enabling load balancing and failover. The disks are presented from a storage via FC to both servers and from these servers the same disks are accessed using the same drive or mount path.Does the Transport Type in Share Mount Path have to be Regular on both servers?Or, on the first server, is it configured as a local path and shared with the second as a Data Server - FC?What is the recommendation according to best practices? 

Hi team,I would like to confirm a specific point regarding Commvault Deduplication.Is it supported to store the Deduplication Database on a NAS , considering that the NAS itself is not a MediaAgent?I want to confirm whether the DDB can reside on NAS storage or if it must be located on MediaAgent-local storage.Thank you. 

Hey Community,A couple of days ago, I received an email with this anomaly notification “Garbage Collection Delay in DDB Stores.” What does this issue indicate, and what should I be checking when this appears?When I click the link to see the more details about this notification, it opens a DDB Pruning Performance Anomaly Report, but it doesn’t seem to provide any useful information. Has anyone run into this before or have suggestions what to do with this?Thanks in advance!