Big news! The community will be moving to a new platform April 21. Read more.
Big news! The community will be moving to a new platform April 21. Read more.
Absent Member.
Absent Member.
3913 views

Data Protector 7.00 - Slow VMware restore to vSphere 5 Environment

Jump to solution

Hey guys and gals,

I currently do full backups of my vmware environment. The backups use transport method SAN and they seem to be pretty snappy. My vCenter is located on the same piece of physical hardware as my Data Protector cell. When I go to do a restore, I see some very peculiar activities. My NICs responsible for SAN communication consistantly spike between 0 and 7% use on a pretty regular interval. 15 seconds at 7% usage followed by 30 seconds or so of 0% usage. When I look in vCenter, I see the message that the system is "Allocating Blocks." It roughly takes me 2 to 3 hours to backup a 600 gig plus environment, but it's taking me an entire hour to restore a single 10 gig thin provisioned ESX server.

0 Likes
1 Solution

Accepted Solutions
Absent Member.
Absent Member.

So, anyway, after searching for quite some time about this issue, it turns out it has to do with thin provisioning and lazy zero writes. It's actually the way VMware is dealing with the restore process, not data protector. For those of you interested, VMware Knowledge Base Article 1035096 references the specific situation where you try to restore a thinly provisioned VM back to a VMFS volume through SAN based transport. 


I am sure someone knew about this before me... 

View solution in original post

2 Replies
Absent Member.
Absent Member.

So, anyway, after searching for quite some time about this issue, it turns out it has to do with thin provisioning and lazy zero writes. It's actually the way VMware is dealing with the restore process, not data protector. For those of you interested, VMware Knowledge Base Article 1035096 references the specific situation where you try to restore a thinly provisioned VM back to a VMFS volume through SAN based transport. 


I am sure someone knew about this before me... 

View solution in original post

Absent Member.. Absent Member..
Absent Member..

Hi there !

I had the same problems with slow restores !

VMware vImage SAN-Backups were over 100Mb/s, but restores ( I used thin disks !) crawled at about 15Mb/s.

 

Switching to NBD AND setting this Parameters in omnirc improved restore-speed to about 50-60Mb/s ,

which is still about half of backup-speed, but at least useable !

 

I made the following Settings to omnirc on the MA-Host ( actually I have the same settings on ALL Hosts !) :

 

### usefull Settings for DP 7.01, Christian  Aichmair

#===========================================================================

# Installation Options on CM and IS #=========================================================================== OB2FWPASSTHRU=1

OB2_SSH_ENABLED=1

 

# Settings for MA :

#=========================================================================== OB2MAREADRETRY=5

OB2MAWRITERETRY=5

OB2FILELIBDISKFULL=1

 

#===========================================================================

# Options to disable Directory-Treewalk during FileSystem-Backups #=========================================================================== OB2NOTREEWALK=1

 

#===========================================================================

# Settings to avoid BAR-Timeouts with DP 7.01 : #===========================================================================

OB2SHMIPC=0

OB2IPCKEEPALIVE=1

OB2INETTIMEOUT=20

OB2IPCKEEPALIVETIME=600

OB2IPCKEEPALIVEINTERVAL=60

OB2RECONNECT_RETRY=900

 

#===========================================================================

# Settings to avoid VEAgent-Timeouts with DP 7.01 and imporve Speed :

#=========================================================================== OB2_VEAGENT_VCENTER_CONNECTION_LIMIT=25

OB2_VEAGENT_VDDK_RETRY_COUNTER=15

OB2_VEAGENT_BACKUP_DISK_BUFFER_SIZE=16

OB2_VEAGENT_RESTORE_DISK_BUFFER_SIZE=16

OB2_VEAGENT_THREADED_BACKUP=1

OB2_VEPA_SEEK_ITEMS=1

 

# use NDB for VMware-Restores !!!

OB2_VEAGENT_RESTORE_TRANSPORT_METHOD=NBD

 

# SAN-Restore is good only for THICK-Disks, and even then not much faster than NBD during restores !!!

# OB2_VEAGENT_RESTORE_TRANSPORT_METHOD=SAN

 

# not sure if the next 2 are even used, but theres no error ;-))

OB2_VEAGENT_RESTORE_BUFFER_SIZE=16

OB2_VEAGENT_NBD_BUFFER_SIZE=16

 

#===========================================================================

# Settings to improve SQL-Performance : #=========================================================================== OB2SQLBLOCKSIZE=65536

 

#######################

 

Hope it helps others to improve reliability and speed,

 

Christian Aichmair

 

 

0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.