Solaris host level SAN migration from Clariion to VMAX – Hands on Lab

Ramdev

Ramdev

I have started unixadminschool.com ( aka gurkulindia.com) in 2009 as my own personal reference blog, and later sometime i have realized that my leanings might be helpful for other unixadmins if I manage my knowledge-base in more user friendly format. And the result is today's' unixadminschool.com. You can connect me at - https://www.linkedin.com/in/unixadminschool/

Loading Facebook Comments ...

34 Responses

  1. Avatar Mahesh R M says:

    Hi,

    Great notes…………………

  2. Avatar Robert says:

    Check out “EMC Open Replicator”.

  3. Ravi Ravi says:

    @Robert: we know there are storage array based migration tools and OR is one of the best, but this article discusses about the Host based migration. I think this is one of the well organized article on the web. Please read the 1st paragraph of this article.

  4. Avatar Harsha says:

    I certainly recommend EMC open migrator for tasks like these- its a host-based migration tool (unlike open replicator). It lets you:

    1. mirror from old set of LUNs to the new without I/O downtime (the only downtime is when you sync the delta).

    2. You can throttle the I/O during the mirroring process – which you can’t via SVM or Tar – this is important if you are copying large data (say > TB) and don’t want to impact the system during business hours.

    3. You can control (pause/resume/restart) multiple mirroring sessions.

    4. Lets you “compare” the data and it will carry over all the permissions as-is.

  5. Avatar Nuno Pacheco says:

    Sometimes, activating “storage software” can be expensive… 
    With SVM you can just set all metadevices as two way mirrors. Normally they should be a one way mirror, then, when in migration times, add the new sub-mirror (new storage array), wait for “sync” and then remove the old sub-mirror (old storage array).

  6. Avatar Ajay says:

    What about mirroring with filesystems are created with SVM soft partitions??? I was trying to mirror and break the mirror for copying data but I guess it has some limitations as it is a not having a physical allocated device.

    • Ramdev Ramdev says:

      Hi Ajay, Yes, the SVM mirroring with soft partitions should have an extra step after you break the mirror
      i.e. copying the exntent headers ( used by soft partitions ) to the new disk slice contains soft partitions, using metarecover

      metarecover -p -m
      where cxtxdxsx — should be the slice ( with soft partitions ) on the detached disk.

  7. Avatar Vishal Shroff says:

    Ramdev I have seen lots of ur post on Linkedin also.
    Excellent knowledge sharing stuff u r doing to all SA’s.

  8. Avatar chetantare09 says:

    Dear Sir,

    We are performing jumpstart on V890 server bot we are getting following error

    {7} ok boot /pci@8,700000/pci@5/pci@0/network@0,0 – install
    Boot device: /pci@8,700000/pci@5/pci@0/network@0,0 File and args: – install
    /pci@8,700000/pci@5/pci@0/network@0: 1000 Mbps full duplex link up
    Requesting Internet Address for 0:14:4f:24:f4:10
    Requesting Internet Address for 0:14:4f:24:f4:10
    /pci@8,700000/pci@5/pci@0/network@0: 1000 Mbps full duplex link up
    ramdisk-root ufs-file-system
    Loading: /platform/SUNW,Sun-Fire-V890/kernel/sparcv9/unix
    Loading: /platform/sun4u/kernel/sparcv9/unix
    SunOS Release 5.10 Version Generic_147440-01 64-bit
    Copyright (c) 1983, 2011, Oracle and/or its affiliates. All rights reserved.
    os-io Configuring devices.
    Using RPC Bootparams for network configuration information.
    Attempting to configure interface eri0…
    Skipped interface eri0
    Attempting to configure interface ge0…
    SUNW,pci-gem0: Could not Establish Link — Check if the cable is plugged in.
    Skipped interface ge0
    Attempting to configure interface ce3…
    SUNW,pci-gem0: Could not Establish Link — Check if the cable is plugged in.
    Skipped interface ce3
    Attempting to configure interface ce2…
    Skipped interface ce2
    Attempting to configure interface ce1…
    Skipped interface ce1
    Attempting to configure interface ce0…
    Configured interface ce0
    |

    ~>/

    rsc>
    rsc>
    rsc>
    rsc>
    rsc> break

    it’s canot going ahead after “configure interface ce0”

    # dladm show-dev
    ce0 link: up speed: 1000 Mbps duplex: full
    ce1 link: down speed: 0 Mbps duplex: unknown
    ce2 link: down speed: 0 Mbps duplex: unknown
    ce3 link: down speed: 0 Mbps duplex: unknown
    ge0 link: down speed: 0 Mbps duplex: unknown
    eri0 link: down speed: 0 Mbps duplex: unknown

    # dladm show-link
    ce0 type: legacy mtu: 1500 device: ce0
    ce1 type: legacy mtu: 1500 device: ce1
    ce2 type: legacy mtu: 1500 device: ce2
    ce3 type: legacy mtu: 1500 device: ce3
    ge0 type: legacy mtu: 1500 device: ge0
    eri0 type: legacy mtu: 1500 device: eri0

    pls help me to sloved this i unable to understand what is promb?????

  9. Avatar Nikhil09 says:

    Hi,
    I am new buddy in SOLARIS… i read book & search in Google also i didn’t get my satisfy answer …….

    In ifconfig -a cmd we recv following o/p

    lo0: flags=2001000849 mtu 8232 index 1
    inet 127.0.0.1 netmask ff000000
    ce0: flags=1000843 mtu 1500 index 2
    inet 10.35.54.21 netmask ffffff00 broadcast 10.35.54.255
    ether 0:94:4f:4a:f7:27

    whats is broadcast address, from which file it will received that & what is roll of that in solaris n/w part???
    also in “/etc/netmasks” file we mentioned server IP range & netmasks value
    whats is role of netmasks ????when mentioned 255.255.255.0 means what happed in background…

    If u see ifconfig -a o/p very carefully every NIC card there is flags is there & after that one number is present “2001000849” & “1000843” whats is indicating this number…

    When OS get install, by default one address is already present that is “lo0”.
    To get access of server we req. IP address & for IP we req. NIC card, so lo0 is SUN H/W by default NIC card??? in that description it’s write the LOOPBACK whats is that “LOOPBACK” whats is exactly task in SOLARIS. & in situation we use this loopback address “lo0” …..

    I hope u under stand my que. & u will help me to sloved this BASIC que.

    • Ramdev Ramdev says:

      Nikhil, your questions are not specific to solaris, they are tcp/ip concepts and you need to search for tcp/ip basics to understand this.

      >>> whats is broadcast address, from which file it will received that & what is roll of that in solaris n/w part???
      >>> Every System as IP Address ( just like every mobile had a number), which consists two parts – network identification part ( just like our BSNL stats with 94, airtel with 99 …etc) + host identification part ( just like our last digits of mobile number).

      in your ip 10.25.54.21 — 10.25.54 will be treated as network address , where as .21 treated as host address. Just in case if you want to send some information to all the hosts in your network i.e. 10.25.54, instead of sending them from each host you just send it to 10.25.54.255 so that it will be received by all the hosts in your network. And this special address is called broadcast address.

      >>> also in “/etc/netmasks” file we mentioned server IP range & netmasks value
      >>> whats is role of netmasks ????when mentioned 255.255.255.0 means what happed in background…

      TCP/IP actually classifies it’s network into multiple classes like A,B,C,D(multicast). And subnetmasks are used to differentiate one class of IP to another IP. So that whenever you broadcast something it won’t go outside of that network range. Please refer this simple tutorial to understand more about this http://youtu.be/QoQmv2VNuX0

      >>> If u see ifconfig -a o/p very carefully every NIC card there is flags is there & after that one number is present “2001000849″ & “1000843″ whats is indicating this number…
      each number in the flag has a meaning, and you can check that from below list.

      0x00000001 /* interface is up */
      0x00000002 /* broadcast address valid */
      0x00000004 /* turn on debugging */
      0x00000008 /* is a loopback net */
      0x00000010 /* interface is point-to-point link */
      0x00000020 /* avoid use of trailers */
      0x00000040 /* resources allocated */
      0x00000080 /* no address resolution protocol */
      0x00000100 /* receive all packets */
      0x00000200 /* receive all multicast packets */
      0x00000400 /* protocol code on board */
      0x00000800 /* supports multicast */
      0x00001000 /* multicast using broadcast address */
      0x00002000 /* non-unique address */
      0x00004000 /* DHCP controls this interface */
      0x00008000 /* do not advertise */
      0x00010000 /* Do not transmit packets */
      0x00020000 /* No address – just on-link subnet */
      0x00040000 /* interface address deprecated */
      0x00080000 /* address from stateless addrconf */
      0x00100000 /* router on this interface */
      0x00200000 /* No NUD on this interface */
      0x00400000 /* Anycast address */
      0x00800000 /* Do not exchange routing info */
      0x01000000 /* IPv4 interface */
      0x02000000 /* IPv6 interface */
      0x04000000 /* Mobile IP controls this interface */
      0x08000000 /* Don’t failover on NIC failure */
      0x10000000 /* NIC has failed */
      0x20000000 /* Standby NIC to be used on failures */
      0x40000000 /* Standby active or not ? */
      0x80000000 /* NIC has been offlined */

      example: if the flags value is 1000849, that means the interface card is:
      0x00000008 + 0x00000001 = is a loopback and the card is up
      0x00000040 = resources are allocated
      0x00000800 = card supports multicasting

      >>> When OS get install, by default one address is already present that is “lo0″.
      >>> To get access of server we req. IP address & for IP we req. NIC card, so lo0 is SUN H/W by default NIC card??? in that description it’s write the LOOPBACK whats is that “LOOPBACK” whats is exactly task in SOLARIS. & in situation we use this loopback address “lo0″ …..

      lo0 is a virtual interface used for tcp/ip own internal testing by assigning it to loopback address i.e. 127.0.0.1 . And this interface not connected to any external network, so you cannot use to connect to the server.

  10. Avatar Nikhil09 says:

    Hi sir,

    Many many thanks 4 in depth explanation. now i have clear idea about this.
    But yesterday we add to one IP addresss in /etc/hosts

    10.25.54.34 TEST3 TEST3..com localhost.

    whats is diff bet loghosts & localhost
    In which situation we will loghosts & localhost?????

    Again thnks 4 explanation….

  11. Avatar pawan04 says:

    Hi,

    I have perform IPMP on server. but its showing following errors,

    root@DEVTEST10 # dladm show-dev
    ce0 link: up speed: 100 Mbps duplex: full
    ce1 link: up speed: 100 Mbps duplex: full
    root@DEVTEST10 #

    root@DEVTEST10 # ifconfig -a
    lo0: flags=2001000849 mtu 8232 index 1
    inet 127.0.0.1 netmask ff000000
    ce0: flags=9040843 mt u 1500 index 4
    inet 10.45.23.21 netmask ffffff00 broadcast 10.45.23.255
    groupname accgrp
    ether 0:14:4f:4a:c7:37
    ce0:1: flags=1000843 mtu 1500 index 4
    inet 10.45.23.20 netmask ffffff00 broadcast 10.45.23.255
    ce1: flags=9040843 mt u 1500 index 3
    inet 10.45.23.22 netmask ffffff00 broadcast 10.45.23.255
    groupname accgrp
    ether 0:14:4f:4a:c7:36

    cat /etc/hosts
    10.45.23.20 DEVTEST10 DEVTEST10..com loghost
    10.45.23.21 DEVTEST10-ce0
    10.45.23.22 DEVTEST10-ce1

    root@DEVTEST10 # cat /etc/hostname.ce0
    DEVTEST10-ce0 netmask + broadcast + deprecated -failover group accgrp addif DEVTEST10 netmask + broadcast + up
    root@DEVTEST10 # cat /etc/hostname.ce1
    DEVTEST10-ce1 netmask + broadcast + deprecated -failover group accgrp
    root@DEVTEST10 #
    root@DEVTEST10 # if_mpadm -d ce0
    connect: Connection refused
    bind: Cannot assign requested address
    Cannot establish communication with in.mpathd.

    I does not get what wrong i have done that why it’s not done….

    One more thing in /etc/hostname.ce0 file if gave “+failover” same in /etc/hostname.ce1 file also thn whats happended????

  12. Avatar mohammed mukram says:

    Hi Ramdev. Nice Website, which has effective articles about Solaris.. Can u post a realtime implemented document for STICKY BIT operation. where I want to create a file for the user and provision sticky bit rules to any one of the file . 1. Create a user , 2. create a file by the same user, 3. give Sticky bit modification to that file. 4. delete the file by the same user .

  13. Avatar Siva says:

    Hi Ram,

    We have a plan to migrate our old SAN to new one in our Solaris platform with Zones. Please share are the prechecks need to be done before performing migration.

    Regards,
    Siva

  14. Ramdev Ramdev says:

    Hi Siva,  I will check and share if I have something useful for you

  15. Avatar ravi says:

    hi ram
    cd /var   ——> i am getting permissions denied
    and  also cd / —-> same
    what is the problem   

  16. Avatar sanjeeva says:

    Hi Ram,
    i have sys1 and sys2 servers ,i want to migrate the existing zone from sys1 to sys2, i installed zone in sys1 naming as myzone by using /myzone (zonepath) directory.Now i halted the zone and detached the zone and took back of zonepath by using tar command and transported this backedup (zonepath.tar) from sys1 to sys2 by using http://ftp.In sys2 server by using this directory how can i use this backedup directory and can attach the  zone.

    • Ramdev Ramdev says:

      Hi Sanjeeva, Apologies for delayed reply, i am in vacation for past one week and haven’d had chance to look at the questions.

      you just need to extract the backup to new location, and then configure the new zone sys2 using the zonepath pointing to the extracted directory. and then attach the zone to the server. Once attached

  17. Avatar Zaina Ariffin says:

    Hi Ramdev,

    Do you have steps as to how to do sancopy on a solaris 10 machine from a CX4-120 to VNX5500..I am doing migration. Thanks

  18. Ramdev Ramdev says:

    Hi Zaina, Sorry, I don’t have any ready docs for that configuration.

  19. Avatar Zainal Ariffin says:

    Hi Ramdev,
    What I am trying to say if data migration is done by san copy, do i need to do anything on the solaris 10 os itself like the steps you do above.

  20. Ramdev Ramdev says:

    Hi Zainal, The procedure from os part depends on  how you are currently using the disks. for example if you are using the disks under vxvm you will unmount, and deport the diskgroup  before hte san copy. later when you get new luns you will reddetec luns, and the volumes will be imported back to solaris.

    btw, please let me know how are you using the current san devices on your current machine, so that I can give you some outline of instructions.

  21. Avatar Zaina Ariffin says:

    Hi Ramdev, I am using SVM.

    • Ramdev Ramdev says:

      Hi Zaina, SVM is little bit risky. I am afraid my instructions might not work for you for 100% becuause it has so much limitations on the hardware part. It is all about you keep the meta device configuration safe on the server and reinitialize same configuration on the new device. But that will be possible only if the new device and old device having same geometry otherwise it might not work the way you expect and breaks the volume.

      the problem with SVM is the Volume information doesn’t storage on the disks unlike VxVM but they store in the machine.

      If the SAN level copy happening same kind of SAN devices then the below procedure can work:
      Just keep the copies safe before migration:
      A current or accurate copy of metastat -p; metatadb -i output
      A current or accurate copy of the md.cf file
      An up-to-date md.tab file

      commentout the the /etc/vfstab for SAN volumes for not load until the migraiont, unmount the filesystems. and go for san copy.

      And once the new SAN device detected, makesure the device detected as same device path as earlier, if not remember the new device . And initiate the metadevice with exactly same size as earlier metadevice.

  22. Avatar Zaina Ariffin says:

    Thanks Ramdev. I have to initiate new metadevice as it generated new PP devices. I did a host-based copy after that.
    Perhaps you can create a new technical guide on the Solaris for VxVM, ZFS and UFS when there’s data migration from one EMC storage to another. That will be useful for system admin out there. Thanks again.

  23. Ramdev Ramdev says:

    Hi Zaina,  Thanks for updating me the status.

  1. September 16, 2015

    […] Read – Solaris host level SAN migration from Clariion to VMAX – Hands on Lab […]

  2. September 18, 2015

    […] Read –  Solaris host level SAN migration from Clariion to VMAX […]

  3. September 22, 2015

    […] Read – Solaris host level SAN migration from Clariion to VMAX […]

What is in your mind, about this post ? Leave a Reply

Follow

Get every new post delivered to your Inbox

Join other followers