β03-22-2016 07:02 AM
Hi,
Our Environment:
Master Server : 7.7 on WIn 2012 Server
Media Server : Appliance 5230 on 7.6.0.4
We are backing up almopst 400 Servers and we have around 40 SQL Servers and rest are File Servers
The Backups are working fine
We need to find out if our Vmware Policies are able to take maximum advantage of Deduplication. Our MS Windows Policies gives us a DeDup Rate of 95% against our VMware Policies which is around 85%. Is that the Best that we can get or can we still get more out from Vmware Policies. Please check the OPsCentre Report which shows Dedup Data of Feb and for March.
Deduplication Rates by Policy Type | ||||
Feb 1, 2016 12:00:00 AM to Feb 29, 2016 11:59:00 PM (Custom) | ||||
Pre Dedup Size(TB) | Post Dedup Size(TB) | Deduplication Rate | Job Size Deduplication Savings(TB) | |
MS-SQL-Server | 7.24 | 1.32 | 81.79 | 5.92 |
MS-Windows | 48.13 | 2.85 | 94.09 | 45.28 |
Standard | 2.66 | 0.03 | 98.68 | 2.62 |
VMware | 345.98 | 53.6 | 84.51 | 292.37 |
Deduplication Rates by Policy Type | ||||
Mar 1, 2016 12:00:00 AM to Mar 22, 2016 9:47:28 AM (Custom) | ||||
Pre Dedup Size(TB) | Post Dedup Size(TB) | Deduplication Rate | Job Size Deduplication Savings(TB) | |
MS-SQL-Server | 1.73 | 0.04 | 97.89 | 1.7 |
MS-Windows | 30.05 | 0.86 | 97.14 | 29.19 |
Standard | 1.9 | 0.03 | 98.53 | 1.87 |
VMware | 252.29 | 34.82 | 86.2 | 217.47 |
Solved! Go to Solution.
β03-23-2016 12:47 AM
There are some bad dedupe client in-between. The only option is to re-direct backup of those databases (or VMs) to non-dedupe storage. Alternative just ignore and use the MSDP space required. I would recommend re-directing backup of those VM's to non-dedupe storage
As mentioned before, this dilemma is not unknown in a dedupe world. It is just hitting you now, but good work you discovered it before you ran of of MSDP disk space !!
It seem all the _BATCH* suffer from bad dedupe rates.
β03-22-2016 08:20 AM
Looks OK to me initially, you got the majority of the data in VMware - 345 TB pre dedupe, stored in 53 TB. Almost 100TB increase from previous month
But as always be careful what people put inside the VM's. Encrypted data or precompressed data is poison for dedupe. You need to monitor each VMware job and take action if some jobs show really awful deduplication ratios (remember the dedupe ration in the activity monitor). This cat after mouse chase is no diffrent than having e.g Data Domains around.
There is no setting in Netbackup that will increase dedupe ration by itself - it's all based on what type of data you put into the MSDP pool.
Also remeber that dedupe ration may increase over time when data is being backup repeatedly.
β03-22-2016 12:39 PM
Thanks for your suggestion Nicolai.
We did again have a look at our Dedup Rate for Clients and it seems we are having issues getting a Good rate on Servers which has SQL Database.
These are Sharepoint Server Database but we just backup them up using Vmware Policy and selecting Application State Capture and not using the Sharepoint Policy.
Please look at this Dedup rate for SLQ Servers
Job ID | Job Policy | Job Schedule | Client | Kilobytes | KB/Sec | Elapsed Time | Media Server | Deduplication Rate |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_QA | Full | Client | 1914491744 | 65852 | 8:11:06 | Appliance | 52% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_DEV | Full | Client | 680921760 | 57217 | 3:21:05 | Appliance | 72.20% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH1 | Daily | Client | 90004416 | 24329 | 1:02:24 | Appliance | 31.10% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH2 | Daily | Client | 79931712 | 26240 | 0:51:29 | Appliance | 34.60% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_DEV | Full | Client | 681031968 | 63961 | 2:59:27 | Appliance | 59% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_QA | Full | Client | 1919030848 | 90274 | 6:00:58 | Appliance | 59.80% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_DEV | Full | Client | 681076992 | 61477 | 3:07:06 | Appliance | 59.60% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH1 | Daily | Client | 282380576 | 53814 | 1:29:21 | Appliance | 20.70% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_QA | Full | Cient | 1947966208 | 85117 | 6:33:31 | Appliance | 51.60% |
Backup | PTC-VMWARE-BRONZE-SHAREPOINT_SQL_DEV | Full | Client | 681076896 | 72115 | 2:39:47 | Appliance | 66.40% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH1 | Daily | Client | 197884544 | 27318 | 2:22:23 | Appliance | 19.20% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH3 | Daily | Client | 5076800 | 3344 | 0:28:17 | Appliance | 53.40% |
Backup | PTC-VMWARE-BRONZE-60DAY_BATCH3 | Daily | Client | 20672352 | 8900 | 0:41:05 | Appliance | 30.80% |
β03-23-2016 12:47 AM
There are some bad dedupe client in-between. The only option is to re-direct backup of those databases (or VMs) to non-dedupe storage. Alternative just ignore and use the MSDP space required. I would recommend re-directing backup of those VM's to non-dedupe storage
As mentioned before, this dilemma is not unknown in a dedupe world. It is just hitting you now, but good work you discovered it before you ran of of MSDP disk space !!
It seem all the _BATCH* suffer from bad dedupe rates.
β03-25-2016 02:00 PM
Thanks Nicolai, I will do some more digging about SQL Database Dedup Rates