Skip to main content
Answer

Mount paths: Several smaller ones or fewer large ones? Max (practical) size of them

  • August 5, 2025
  • 4 replies
  • 81 views

Forum|alt.badge.img+11

Background: We have 2 media agents at a secondary site that each have a single “data” mount path to storage that we just hit 256 TB on (OS refuses to expand them further, its using 64kb blocks).  The media agents have their storage/mount path mapped to that dedicated storage volume (like X:\) and then they also map (in commvault) to the other media agent.  So each media agent can read/write to either X:\ on either media agent.

Note: All the volumes in commvault are pointing to the same dedicated storage array (that only commvault uses)

Questions: Since we cannot expand the existing mount path storage beyond 256 TB for either mount, a few options are being explored:

  1. Make a new volume and format it to 128 kb blocks, so it can grow to 512 TB is needed. Then do a “move mount path” of the 256 TB bound volume to the 512TB bound one. result: we have a single mount point per media agent, and it just is really big. using 128 kb blocks seems like a “probably don’t want to do” from some online things I have read about Windows performance of these. Why do the move? I missed the max was going to be 256 TB so there’s quite a bit of TB’s we allocated we cannot use, so it would be to free up all the extra TB’s when the old mount point was deleted.
  2. Make several new volumes (say Y:\, Z:\) and have each be “smaller” (like 64 or 128 TB). Add the mount points and just let commvault use them as it sees fit. leave the large 256 TB mount point alone, as adding more mount points will allow growth.
  3. ???

Basically: I have seen some posts about this here, and it seems like having a single large mount point is kinda a bad thing vs having many small (dedicated) ones.  I’m wondering if there is a recommended “max mount point size” for “practical” reasons (not system limits!) for management/administration/performance purposes The only obvious one is “its easier to move them as they are smaller” and “if something happens to one, its less of a disaster vs losing the one giant one”.

 

 

Best answer by sbhatia

Hi ​@tigger2 , 

Let me try, 

there’s no hard Commvault software limit on mount path size, but from my admin experience, it’s smart to keep individual mount paths in the 64–128TB range. This sizing strikes the right balance,  it minimizes blast radius if a path fails or needs maintenance, allows software to efficiently parallelize jobs across multiple paths for better performance, and makes day-to-day management (growth, troubleshooting, or moving data) much more straightforward. Stretching a single mount point to be as large as the OS/filesystem allows increases operational risk and administrative pain, so best practice is always to spread your storage across several reasonably sized, dedicated mount points. 
I hope this is what you were looking for. 

4 replies

sbhatia
Vaulter
Forum|alt.badge.img+9
  • Vaulter
  • Answer
  • August 6, 2025

Hi ​@tigger2 , 

Let me try, 

there’s no hard Commvault software limit on mount path size, but from my admin experience, it’s smart to keep individual mount paths in the 64–128TB range. This sizing strikes the right balance,  it minimizes blast radius if a path fails or needs maintenance, allows software to efficiently parallelize jobs across multiple paths for better performance, and makes day-to-day management (growth, troubleshooting, or moving data) much more straightforward. Stretching a single mount point to be as large as the OS/filesystem allows increases operational risk and administrative pain, so best practice is always to spread your storage across several reasonably sized, dedicated mount points. 
I hope this is what you were looking for. 


Forum|alt.badge.img+11
  • Author
  • Byte
  • August 6, 2025

Hi ​@tigger2 , 

Let me try, 

there’s no hard Commvault software limit on mount path size, but from my admin experience, it’s smart to keep individual mount paths in the 64–128TB range. This sizing strikes the right balance,  it minimizes blast radius if a path fails or needs maintenance, allows software to efficiently parallelize jobs across multiple paths for better performance, and makes day-to-day management (growth, troubleshooting, or moving data) much more straightforward. Stretching a single mount point to be as large as the OS/filesystem allows increases operational risk and administrative pain, so best practice is always to spread your storage across several reasonably sized, dedicated mount points. 
I hope this is what you were looking for. 

Thanks! I have another question if you have time/know:

  • Do you know if the mount points “work best” if they are all the same size, or is it ok if they are all different sizes? Like “yes it will technically work” if they are all different sizes but I’m thinking more of “best practice” or “most performant” or “if there can be xyz issues if they are all different sizes”.  Currently we have always tried (when expanding storage) to add the same amount to all our mount points so they are all the same size. not really for any other reason other than to keep them all identically sized.

sbhatia
Vaulter
Forum|alt.badge.img+9
  • Vaulter
  • August 7, 2025

There’s absolutely no problem if your mount points are all different sizes, it works just fine and won’t affect performance or reliability. Still, when all your mount points are similar in size, it becomes much easier to plan capacity, spot issues before they happen, and keep monitoring straightforward, since you don’t have to keep track of which path might fill up first.

By default, Commvault writes to the path with the most free space, so if you have a larger and a smaller mount point, the bigger one will fill a bit faster, but the system will automatically move on to others when one gets full.

So, unless your team specifically wants the simplicity of everything matching up (which does make admin life more predictable), having different-sized mounts is totally fine and won’t cause any trouble, especially as your storage expands in different chunks.


Forum|alt.badge.img+3
  • Commvault Certified Expert
  • August 7, 2025

@sbhatia covered it in great detail.

The only thing I would add to make it more clear on that point is that having one large mount path also creates one large point of failure. If all of your data is on a 512 TB volume and it goes down, you’ve lost access to all of that data. If you have four 128 TB volumes and one goes down, you theoretically have access to some of the data.