TSM for VE performance (slow backup and retore)

t.Dahl

Active Newcomer
Joined
Sep 17, 2012
Messages
13
Reaction score
0
Points
0
Hello,

my backup is very slow and i don't find a problem.

My environment:

Server:Windows 2008 R2 TSM 6.2.5
Client: Windows 2008 R2 TSM for VE 6.4

Between the client and the Data Mover San I put an 8 GByte / s and between the Data Mover and the TSM server is a 10 GByte / s LAN.
I have already made ​​in the network performance test and get to about 800 Mbyte / s However, when I TSM Sure I only get about 100 MByte / s
The VCenter is a Virtual Server.
Here are my times dsm.opt and server settings.


I hope someone can help me or ever had the problem and can give me advice.


dsmc.exe -optfile=dsm.opt -asnodename=DEHZE01-VDC001

dsm.opt
NODENAME dehze01-wis100_DM
TCPSERVERADDRESS 10.255.156.254
PASSWORDACCESS GENERATE
VMVSTORTRANSPORT "SAN"
VMCHOST 10.220.0.40
VMCUSER rhs\srv.dehze01.tsm
REVOKEREMOTEACCESS NONE
MANAGEDSERVICES WEBCLIENT SCHEDULE
VMMC "MB__NL_030_001_060"
VMCTLMC "MC__NL_030_001_060"
SCHEDMODE PROMPTED
SCHEDLOGNAME "C:\Program Files\Tivoli\TSM\baclient\dsmsched_Windows.log"
SCHEDLOGRETENTION 3 D
ERRORLOGNAME "C:\Program Files\Tivoli\TSM\baclient\dsmerror_Windows.log"
ERRORLOGRETENTION 3 D
vmmaxparallel 50
vmlimitperdatastore 0
vmlimitperhost 2
DOMAIN.VMFULL "VMHOST=dehze01-evs021;-VM=*lsv*nagios;"
resourceutilization 10

server:
Target Node Agent Node
--------------- ---------------------------------------------
VC_DEHZE01-WIS- VMCLI
100
VC_DEHZE01-WSV- VM_CLI DEHZE01-VDC001
042
DEHZE01-VDC001 VM_CLI DEHZE01-WSV530_DM DEHZE01-WSV533_DM
DEHZE01-WSV534_DM DEHZE01-WSV535_DM
DEHZE01-WSV537_DM DEHZE01-WSV536_DM
DEHZE01-WSV538_DM DEHZE01-WSV539_DM
DEHZE01-WSV544_DM DEHZE01-WSV543_DM
DEHZE01-WSV542_DM DEHZE01-WSV541_DM
DEHZE01-WSV540_DM DEHZE01-WSV547_DM
DEHZE01-WSV548_DM DEHZE01-WIS100_DM
DEHZE01-WSV582_DM DEHZE01-WSV580_DM
DEHZE01-WSV583_DM DEHZE01-WSV581_DM
DEHZE01-WSV529_DM

q mgmt
Policy Policy Mgmt Default Description
Domain Set Name Class Mgmt
Name Name Class ?
--------- --------- --------- --------- ------------------------
VMWARE ACTIVE MB__NL_0- Yes INCREMENTAL BACKUP,
30_001_- retpd = 30 days
060
VMWARE ACTIVE MC__NL_0- No Control Data BACKUP,
30_001_- retpd = 30 days
060


q copyg
Policy Policy Mgmt Copy Versions Versions Retain Retain
Domain Set Name Class Group Data Data Extra Only
Name Name Name Exists Deleted Versions Version
--------- --------- --------- --------- -------- -------- -------- -------
VMWARE ACTIVE MB__NL_0- STANDARD No Limit 1 30 60
30_001_-
060
VMWARE ACTIVE MC__NL_0- STANDARD No Limit 1 30 60
30_001_-
060



q stg

Storage Device Estimated Pct Pct High Low Next Stora-
Pool Name Class Name Capacity Util Migr Mig Mig ge Pool
Pct Pct
----------- ---------- ---------- ----- ----- ---- --- -----------
FDC2_CONTR- DEVFDC2 2,000 G 0.8 0.8 100 0
OL_VMWARE
FDC2_FILE_- DEVFDC2 5,994 G 52.6 52.6 80 0 LTOPOOL1
VMWARE

q devc

Device Device Storage Device Format Est/Max Mount
Class Access Pool Type Capacity Limit
Name Strategy Count (MB)
--------- ---------- ------- --------- ------ -------- ------
DEVFDC2 Sequential 2 FILE DRIVE 20,480.0 300

Many thanks in advance
 
In the moment the Storage is connect with Fiber Channel to the ESX Server. In had test it with SAN between vmdk and Datamover i only get 120 -130 Mbyte/s.
In the next step i will test backup with a Storage Agent.
 
You are currently sending both data and control files data to a sequential FILE type of device class in storage pool . You should have a dedicated random access DISK type of storage pool for the control files which you point to using the VMCTLMC option in the dsm.opt file. The control files should NEVER migrate off of the Disk to a Sequential medium especially physical tape. I think this will help.

http://pic.dhe.ibm.com/infocenter/t...pic=/com.ibm.itsm.ve.doc/r_ve_mount_tape.html
 
Hi i don't migarte the config files to LTO. The file are safe in the storage disk in Raid5.
 
Hey there.


Checking your OPT file I see you have "vmmaxparallel 50".. I think that number might be a little to high and causing a bottleneck.
I would try 10 or less if you have big VMs. I am not doing more than 5 or 6 VMs at the same time and having good results with it.


Maybe you can start with a low number and getting high if you get the desired performance.


Let us know if that helps.


Thank you!
 
Back
Top