Besser UBER : Celerra VSA UBER v2

**** Go here for new version of UBER VSA ****

What is better than UBER?

UBER version 2.

It has only been 19 days since the original release of the Celerra VSA UBER edition for the masses. The response has been overwhelming and encouraging. And out of that appreciation I am excited to present the Celerra VSA UBER V2 (version 2). It has a host of new features and now features the combined input of the vSpecialist team to put the shine on it.

If you are not familiar with the original feature list with the first UBER version, then go here first.

Here are the new features and enhancements in UBER Version 2:

  1. Even shorter initial setup
  2. Better network configuration. Removal of old style Control Center config from original VSA (Thanks to Eric Hollis & Kevin Z for help)
  3. Completely automated addition of storage. You heard that right… If you want to add more storage simply shutoff the VSA. Add as many hard drives as needed and turn back on. The VSA now detects the new VMDK’s, partitions, formats, mounts, performs Clariion configuration, and adds to storage pools. Gone are the days of manually having to configure. See video below for demo. (This by far was the biggest request)
  4. The VSA now includes a Rapid Configuration script for setting up CIFS, NFS, and Replication between 2 VSA’s. This script will accelerate the creation of a baseline working set of VSA’s. We highly recommend you use this after becoming familiar with the Celerra management first. It is not meant to replace Control Center. (Thanks to Clint Kitson for the single-handed genius of this feature)
  5. Now the standard configuration wizard includes setup of NTP.
  6. Further speed improvements around network latency. Should see more stable and lower access times with NFS and iSCSI.

Couple outstanding issues:

  1. Some setups will get alerts around ‘slot_3’ being reset or in a stale state. You may ignore this error. This slot is for the unused data mover. Still narrowing down why this happens to some setups.
  2. If you go far along in the setup of replication and run into funky issues it is easier to redeploy the VSA and start over. Remember this is a simulation of a set of hardware. Not everything reverses as easily as it deploys. With the new wizards it is quicker this way too.

The new downloads are:

**** Go here for new version of UBER VSA ****

Please leave comments with opinions, questions, suggestions, or favorite recipes. The more comments the more inspired I will be to get the next rev out. Or maybe even something new that is not Celerra…

And while you wait for your download, watch this video showing off the new features(go full screen for more detail):


Celerra VSA Videos

60 Comments Leave a comment

  1. Hello,rnrnIs anyone else running into an issue where new disks are not being added to the Celerra? On first boot after the vmdks have been added the system see the new vmdks, works it’s magic by auto partitioning, but once fully booted and I look in Unisphere at the storage pools, I still see the default 18 and 4 GB Pools and the extra 80 GB is nowhere to be found.rnrnAm I doing something wrong?rnrnThanks,rnJkasal

  2. Great job on making this thing cake to setup. I didn’t think it was totally hard the old way, but this makes it so no TC should have any excuse not to install and have it up in running in minutes. Also, big kudos on the adding storage component.

  3. Great stuff. I wonder if you could add in to your script configuration for eth2, and eth3? It would be nice if the use of those NIC’s agreed with Bernies book. So eth0 for iscsi, eth1 for nfs, and eth2 for mgmt. But in any case, it would be nice to have eth2 / eth3 configured.

    Gerat job and thanks very much for doing this.


    • eth1 & eth2 (there isn’t an eth3 in this one) are configured. There just isn’t an IP address bound to these interfaces for Control Center (which binds to eth0 only). When you setup interfaces in Control Center (under ‘Network’) cge0 = eth1 and cge1 = eth2. These interfaces are used by the data movers only.

      Bernie’s doc is based on incorrect info he received on earlier versions. His SRM stuff is canon though. Binding all three and then having the data movers bind to the same interfaces actually slows down the NFS/iSCSI traffic with interrupts.

  4. Thanks Nick. I would like to route ESX host access, and replication traffic via eth1 or eth2. Should I just use netconfig to give them addresses? Is that all that is required?

    • Routing that to those interfaces, if you bind NFS or iSCSI to them, isn’t a great idea.

      In a real Celerra there are more interfaces because the Data Movers have their own. It is better to focus data traffic (NFS or iSCSI) on their own adapters.

      That said if you were only using one of the two (NFS or iSCSI) then you could use the remaining interface for whatever you want.

      And yes, netconfig -d eth[0,1,2] would be the right way to config.

  5. Any chance you can work with Bernie so his book would help build a VSA that is as fast as yours, and use the same amount of resources? I would rather use his book to build everything, but would like it to to have the UBER VSA. Is that possible somehow? The issue is that I would like to have three networks working, and be able to than use his book to do the replication stuff. But currently his networking is flawed I think, and his version of the vsa is not as flexible as yours. And the UBER forces me to be in a position that is fast and good, but hard to use Bernies book to do the rest of the setup – i.e. networking, replication and SRM.

    Any comments?



  6. Hello,

    Thanks Good work.

    I Just have issue,
    I try to use the VSA on DL385 G1 – 2 X Dual Core Opteron 280 in ESXi 4.
    It’s not working correctly CPU at 100% any time, very very slow.
    I think there is no ADM-V instrucution on this cpu. If someone have best parameter
    I try this but no more success

    For the good new!! I change the hardware to intel Mac 2,5 and it’s working perfectly….
    very quik …

    Thanks for you work


  7. Hi,
    I’ve got an issue. The https access does not work. I can access the mgmt console using telnet.

    I tried the troubleshoot in Appendix D of the doc :
    /nas/http/nas_ezadm/etc/script restart
    But it hang when starting tomcat web server.

    The last lines of catalina.out are :

    Starting service Tomcat-Standalone
    Apache Tomcat/4.1.39
    Prefetch Started
    Jun 2, 2010 2:50:30 AM org.apache.jk.common.ChannelSocket init
    INFO: JK: ajp13 listening on /
    Jun 2, 2010 2:50:30 AM org.apache.jk.server.JkMain start
    INFO: Jk running ID=0 time=0/81 config=null
    Unable to Prefetch for class com.emc.nasui.apl.ListMoverObject
    Unable to Prefetch for class com.emc.nasui.apl.ListVdmObject
    Unable to Prefetch for class com.emc.nasui.apl.ListFilesystemObject
    Unable to Prefetch for class com.emc.nasui.apl.ListMigrationFilesystemObject
    Unable to Prefetch for class com.emc.nasui.apl.ListCheckpointObject
    Unable to Prefetch for class com.emc.nasui.apl.ListVolumeObject
    Unable to Prefetch for class com.emc.nasui.apl.ListShareObject
    [root@celera-mgmt logs]# ps -e |grep http
    [root@celera-mgmt logs]#

  8. I rebooted the appliance, it works now. But yesterday, I rebooted the appliance too without success. Perhaps the script helped.

  9. sounds awesome! I just ran through the init script and am getting some tomcat errors. Things look good through login. I can view the licenses, statistics, system info and cs properties screen, however the tree never builds in the left pane and when i select the top level IP node i get the error below…. Anybody else?

    HTTP Status 400 – Invalid path /systemStatusDisplay_init=true&queryTime=1276118334217 was requested

    type Status report

    message Invalid path /systemStatusDisplay_init=true&queryTime=1276118334217 was requested

    description The request sent by the client was syntactically incorrect (Invalid path /systemStatusDisplay_init=true&queryTime=1276118334217 was requested).
    Apache Tomcat/4.1.39

  10. Nick,

    I am on an AMD opteron setup running windows 7 and worskstation 7.1. I don’t know if this helps, but I thought that I would include it so that you could determine what setups get this. Let me know if you need further details.


  11. Hi,

    I am trying to use your VSA. Need your help . I am using the VSA in a VMWARE workstation 7. I am able to ping VSA from host but not able to access from IE, ending with error Page Cannot be displayed. I am tring to check IP settings in VSA and its asking for password. Plz help me…


  12. Hi Nick,
    I’ve come across a problem with the v2 VSA.
    This only applies to ESX and not workstation.
    In CIFS I get the following when I try to copy a file “Error Copying File or Folder” – “The path is too deep”. The directory copies fine, but only 1 or 2 of the files copy before the error.
    If I create an iSCSI target Windows can’t format it.
    If I mount an NFS share to ESX, i get and IO error when copying files to it.

    If I mount into the folder from the control station via quotas it will let me copy and delete files in no problem.

    V1 worked like a charm for me and at this point I’m all out of ideas to try to fix this.

    Has anyone else had this same problem?
    I’m running ESXi 4.0.0,261974

  13. Hi I am not able to Download Celerra VSA UBER Version 2 – OVA (ESX). Link is not working for me. It began Download for 200 MB or so then showed error folder does’nt exists.

  14. I’m getting ready to build a third ESXi box, which will be a dedicated host for my Celerra VSA.

    Q: Am I on the right track, thinking that I can add as much disk as the mobo supports, and then add oh, say.. 256gb VMDK’s to the Celerra, to produce a VSA with a total capacity somewhat less than physical capacity (due to overhead for the vRAID)?

    Has anyone ever pushed the limits of the EMC VSA? I’m hoping that if all goes well, I can push all of my storage into this thing, at some point ending up with about 10T in my VSA “SAN” at home.. 🙂

  15. I cannot download anything, it keep giving “550 Can't change directory to /vmwarechampion/Virtual_Machine_Library/Celerra/5_6_48_701/UBER/EMC Celerra VSA – v5.6.48.701 – UBERv2.ova: Permission denied” error.

  16. Probably my lack of understanding on this 🙂 but how long does it take to create a file system. I attached 1.5 TB worth of storage in 3 x 512GB VMDK's and tried to create a file system from the auto generated volumes. The task has been going almost a day now with no sign of ending.CheersD

  17. Probably my lack of understanding on this 🙂 but how long does it take to create a file system. I attached 1.5 TB worth of storage in 3 x 512GB VMDK’s and tried to create a file system from the auto generated volumes. The task has been going almost a day now with no sign of ending.rnCheersrnD

  18. When i access the CIFS share i’ve created it’s asking for the username & password.
    I tried using user A/c in the domain controller, but no luck.
    please help.

  19. Hi i´m using the latest Version of Celerra VSA and i am very dissapointet about the Performance. It only come up to approximately 7MB/s with GigaBit Ethernet? Is this an Limitation of the Box, because when i run iperf or dd tests, i get much better results, so it seems to be related to the Simulator Configuration.
    Thanks in advance.

  20. Hi

    Is it possible to use the VTLU funtion in this VM? I have added some scsi tape-devices to the VM, and I can see them from the Linux kernel, but is it possible to use them in the Celerra part?


%d bloggers like this: