PDDO diskpool space free discrepancy

[Ryan_VDW]

Hi All

I have a test system with Netbackup 6.5.4 and Puredisk 6.6.0.2

The Puredisk Content Router has a 2.3TB partition mounted as /Storage/data and df -h reports that 258GB of this is currently available.

From the Puredisk Web GUI - the capacity dashboard shows that the Total Disk space on the content router is 2.21TB, of which 1.93TB is used (87.49%) - although this doesn't quite match up with the information from the Suse Linux console the free space is pretty much the same.

However, from the NetBackup Console, NetBackup reports that the PDDO diskpool is 2.21TB in size with 2.15 TB used and only 56GB available (97.49%) and as a result it will not kick off any backups as it thinks that the disk store is full. (The high water mark is 98%)

It seems as though 10% of the available capacity is being reserved for some reason. Any ideas how to get NetBackup to report the accurate diskpool utilisation?

Thanks

[Mik_Archer]

does this not cover this 10% ?

Out of the Netbackup De-Dupe Guide.

The deduplication database consumes approximately 10 percent of the storage

capacity. Therefore, approximately 90 percent of the storage capacity is usable

space for unique backup data. The actual percentages vary depending on the data.

[Ryan_VDW]

If Netbackup was using 10% for the deduplication database would this not be part of the used space as reported from the linux console of the PureDisk Content Router when using "df -h". What I don't understand is that NetBackup is reporting a used space figure which is 200GB higher than the actual used space figure.

[Kevin_Good]

It could be that NBU is looking at the space used in /Storage (eg du /Storage) and comparing it to the free space in /Storage/data

Either Way, you are right at the limit of what you should use on a content router, and should grow it, or add a second.

[Casey_Simpson]

I noticed the same issue that you are talking about. Everything that Puredisk is showing you is 10% more than Netbackup believes that it sees. The reason for this is that Puredisk by default will not allow the volume to go beyond 90% full, thus reserving 10% of the space to work with. I've been told that the reason for this is that Puredisk engines need free space to work with for routine maintenance tasks. If it runs out of space, you are at high risk of data corruption because it cannot properly process transaction logs and commit data.

That being said, Support did give me a workaround for this to allow me to grow to 95% utilization, however they STRONGLY advised that I not go beyond that due to risk of data corruption and inability to fix if it does not have the space needed to function. What I mean by this statement, is that you can make the same changes that I've done to get an extra 5% out of your Puredisk storage pool, but do not go beyond that or you will risk serious problems.

1) SSH over to your Puredisk node, locate 3 files under /Storage/etc/puredisk, agent.cfg, pdmbe.cfg, and contentrouter.cfg

2) Make a copy of these 3 files before you make any changes.

3) Open up agent.cfg with a text editor. Locate "softLimit", set to 90. Locate "HardLimit", set to 95

4) Open up contentrouter.cfg with a text editor. Set "LowSpaceThreshold" to 5%, and "VeryLowSpaceThreshold" to 3%

5) Open up pdmbe.cfg with a text editor. Set ""warningSpaceThreshold" to 5, and "veryLowSpaceThreshold" to 3

6) restart Puredisk services. Make sure to make this change on all Puredisk servers in the system.

7) Refresh the disk pool in Netbackup, you should now see 5% more space.