Forum Discussion

john_levis's avatar
john_levis
Level 3
27 days ago

Vm backup to disque extremly slow

Hi all,

i have a physical netbackup 10.3 server with Win2016 running  "fine" with vm backup to Lto8 tapes and disk.

I bought a new server to upgrade in win2022 and i installed on it a fresh version in 10.5.

i created a disk in storage unit. added my vcenter in credential and created a policy to test a vm backup.

in attachement the backup log. 

after 01:42:59 of time elapsed the backup is at 39 KB/sec and has backuped 200000 kb of data.

the same Vm backuped by the older Master is at 20867 KB/Sec in an 00:03:25 time.

i tested with other vm and it's the same.

Other problem : when i go in the gui in Policies or Catalog, (not after instal but after policy creation)  i got the message "Could not get Class_att_defs information: Intermittent connectivity lost with the server ....." (in attachement)

What could i check/modify to make this fresh master running fine please ?

 

 

10.5
Intermittent connectivity lost with the server
slow backups
VM backup
  • kiyan's avatar
    kiyan
    Level 6
    26 days ago

    is it possible to provide more information?
    -how old NetBackup is connected to VMware (SAN, hotadd,NBD),is it same for new NetBackup server?
    what is the bandwidth between VMware and NetBackup servers ?
    how is the policy configured in old NetBackup and new NetBackup (classic, Intelligent)
    can you share activity report of one VM backup which is finished in the old and new NetBackup server?
    both of NetBackup servers are going through same security devices?or they have different subnets and rules applied to them?you can check with your network team about routing as well.
    as per the activity job, I can only see that Data at transfer encryption is enabled, is that on purpose ?(by default it is preferred to enabled in the settings).

    • john_levis's avatar
      john_levis
      Level 3
      24 days ago

      the old netbackup is connected to vmware for more than 8 years, the new one has just been connected and share the same ressources for lto tapes/robot and datastores.

      Esxi, Lto, SAN and the Netbackup servers are connected with 10gb lan and 16gb san Fiberchannel.

      the policy are the same for the 2 servers .. i change to local storage to test but was not better.

      the report for the new one is already in attachment. it doesn't finish so i kill it every time. the old report is attached now.

      the 2 masters are on the same network and Vlan with no specific rule.

      The DTE option is the same as the old master and is the default option from installation. i never changed it.

       

       

       

  • john_levis's avatar
    john_levis
    Level 3
    21 days ago

    Anyone any idea please ? i'm dumbfunded that a fresh netbackup installation in not able to backup correctly a Vm on a Vmware cluster.

    • quebek's avatar
      quebek
      Moderator
      20 days ago

      Hi,

      Let's return to the basics checks:

      are your NIC connected on full duplex, with 1 or 10 Gbps?

      are there any routers in between vm, media and primary?

      did you check with nbperfchk the local disk performance?

      https://www.veritas.com/support/en_US/doc/15263389-161340675-0/v125422597-161340675

      Answer what transport mode you are using.

  • john_levis's avatar
    john_levis
    Level 3
    20 days ago

    Hi, thank you for your reply.

    the nic is connected in full duplex 10gbit on the same router/ vlan as vm's esxi.

    network team already checked and the port/vlan parameters for this server is OK.

    here the result for nbperfchk

    D:\Veritas\NetBackup\bin\support>nbperfchk -i e:\NetBackup_10.5_Win.zip  -bs 1024k -o Null
    Statistics log are recorded in nbperfchk_results.log
    read data size is zero, input thread exit!
        2914 MB @  971.0 MB/sec,     2914 MB @  971.0 MB/sec, at 1732284501
    meter thread exit!
        2914 MB @ 2446.7 MB/sec

    The transpot mode is currently LAN for the backup. i marked in the policy first san then hotadd and finally nbd.

    all optimizations boxes are selected.

     

    • quebek's avatar
      quebek
      Moderator
      20 days ago

      Have you tried to backup the other VM? How many files the test one is having?

      Could you disable this optimization  Enable file recovery from VM backup

      it will rule out the mapping of files for that test vm backup.

      Does the new primary/media server see the LUNs of VMWare farm via SAN? If not uncheck SAN. New primary/media is physical you said, so disable HotAdd just leave out there NBD as transport method.