Cannot connect on socket - Veritas Net Backup

This is a discussion on Cannot connect on socket - Veritas Net Backup ; We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem started when we migrated our servers to private IP address. It is not connecting to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: Cannot connect on socket

  1. Cannot connect on socket


    We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem started
    when we migrated our servers to private IP address. It is not connecting
    to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to load
    properties of this client from Host properties in the main console it gives
    the error message "scigate.ncsi.iisc.ernet.in cannot connect on socket".
    The status code is 25.

    We have several other Linux clients. One of these is Red hat Linux 7.1. It
    works fine for all other clients.

    Any help will be appreciated.
    Thanks


  2. Re: Cannot connect on socket


    "Filbert Minj" wrote:
    >
    >We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem started
    >when we migrated our servers to private IP address. It is not connecting
    >to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to load
    >properties of this client from Host properties in the main console it gives
    >the error message "scigate.ncsi.iisc.ernet.in cannot connect on socket".
    >The status code is 25.
    >
    >We have several other Linux clients. One of these is Red hat Linux 7.1.

    It
    >works fine for all other clients.
    >
    >Any help will be appreciated.
    >Thanks
    >


    Filbert,

    The first thing I'd do is ping and traceroute to the Linux box from the
    master server and vice versa. If you can't do that, then NB isn't going
    to work.

    If that's successful, the problem could be that the client is no longer
    able to authenticate the NB server. This happened to me when I tried backing
    up some boxes that had been set up in a psuedo-DMZ: the DNS for the zone
    didn't have my NB servers in its database and therefore the authentication
    failed.

    From the client, cd /usr/openv/netbackup/bin and run

    bpclntcmd -pn

    bpclntcmd -hn media_server_name

    bpclntcmd -ip media_server_IP

    You should get responses back from each one verifying the master server and
    the media server respectively.

    If this doesn't work, you might need to either configure a DNS in this address
    space or add host file entries to resolve the NB server names.


    Scott

  3. Re: Cannot connect on socket


    Scott,

    I tried the steps you have mentioned from the client and I got the responses
    like this:

    [root@scigate bin]# ./bpclntcmd -pn
    expecting response from server ncsi-181
    scigate.ncsi.iisc.ernet.in scigate.ncsi.iisc.ernet.in 10.180.1.14 33474

    [root@scigate bin]# ./bpclntcmd -hn ncsi-181
    host ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    checkhname: aliases:

    [root@scigate bin]# ./bpclntcmd -ip 10.180.13.15
    checkhaddr: host : ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    checkhaddr: aliases:

    Here ncsi-181 and 10.180.13.15 are hostname and ip of media server.
    All these responses look ok.

    One more thing I noticed is when I tried to connect client using telnet
    (telnet 10.180.1.14 13782), it closes the connection immediately where as
    other clients do not. That means client is refusing the connection although
    all the services (process) in clients are running.

    Do have any idea why this is happening?


    wrote:
    >
    >"Filbert Minj" wrote:
    >>
    >>We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem

    started
    >>when we migrated our servers to private IP address. It is not connecting
    >>to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to load
    >>properties of this client from Host properties in the main console it gives
    >>the error message "scigate.ncsi.iisc.ernet.in cannot connect on socket".
    >>The status code is 25.
    >>
    >>We have several other Linux clients. One of these is Red hat Linux 7.1.

    >It
    >>works fine for all other clients.
    >>
    >>Any help will be appreciated.
    >>Thanks
    >>

    >
    >Filbert,
    >
    > The first thing I'd do is ping and traceroute to the Linux box from the
    >master server and vice versa. If you can't do that, then NB isn't going
    >to work.
    >
    > If that's successful, the problem could be that the client is no longer
    >able to authenticate the NB server. This happened to me when I tried backing
    >up some boxes that had been set up in a psuedo-DMZ: the DNS for the zone
    >didn't have my NB servers in its database and therefore the authentication
    >failed.
    >
    >From the client, cd /usr/openv/netbackup/bin and run
    >
    >bpclntcmd -pn
    >
    >bpclntcmd -hn media_server_name
    >
    >bpclntcmd -ip media_server_IP
    >
    >You should get responses back from each one verifying the master server

    and
    >the media server respectively.
    >
    >If this doesn't work, you might need to either configure a DNS in this address
    >space or add host file entries to resolve the NB server names.
    >
    >
    >Scott



  4. Re: Cannot connect on socket


    "Filbert Minj" wrote:
    >
    >Scott,
    >
    >I tried the steps you have mentioned from the client and I got the responses
    >like this:
    >
    >[root@scigate bin]# ./bpclntcmd -pn
    >expecting response from server ncsi-181
    >scigate.ncsi.iisc.ernet.in scigate.ncsi.iisc.ernet.in 10.180.1.14 33474
    >
    >[root@scigate bin]# ./bpclntcmd -hn ncsi-181
    >host ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    >checkhname: aliases:
    >
    >[root@scigate bin]# ./bpclntcmd -ip 10.180.13.15
    >checkhaddr: host : ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    >checkhaddr: aliases:
    >
    >Here ncsi-181 and 10.180.13.15 are hostname and ip of media server.
    >All these responses look ok.
    >
    >One more thing I noticed is when I tried to connect client using telnet


    >(telnet 10.180.1.14 13782), it closes the connection immediately where as
    >other clients do not. That means client is refusing the connection although
    >all the services (process) in clients are running.
    >
    >Do have any idea why this is happening?
    >
    >
    > wrote:
    >>
    >>"Filbert Minj" wrote:
    >>>
    >>>We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem

    >started
    >>>when we migrated our servers to private IP address. It is not connecting
    >>>to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to

    load
    >>>properties of this client from Host properties in the main console it

    gives
    >>>the error message "scigate.ncsi.iisc.ernet.in cannot connect on socket".
    >>>The status code is 25.
    >>>
    >>>We have several other Linux clients. One of these is Red hat Linux 7.1.

    >>It
    >>>works fine for all other clients.
    >>>
    >>>Any help will be appreciated.
    >>>Thanks
    >>>

    >>
    >>Filbert,
    >>
    >> The first thing I'd do is ping and traceroute to the Linux box from

    the
    >>master server and vice versa. If you can't do that, then NB isn't going
    >>to work.
    >>
    >> If that's successful, the problem could be that the client is no longer
    >>able to authenticate the NB server. This happened to me when I tried backing
    >>up some boxes that had been set up in a psuedo-DMZ: the DNS for the zone
    >>didn't have my NB servers in its database and therefore the authentication
    >>failed.
    >>
    >>From the client, cd /usr/openv/netbackup/bin and run
    >>
    >>bpclntcmd -pn
    >>
    >>bpclntcmd -hn media_server_name
    >>
    >>bpclntcmd -ip media_server_IP
    >>
    >>You should get responses back from each one verifying the master server

    >and
    >>the media server respectively.
    >>
    >>If this doesn't work, you might need to either configure a DNS in this

    address
    >>space or add host file entries to resolve the NB server names.
    >>
    >>
    >>Scott

    >

    Filbert,

    I've seen that behavior before [telnet immediately disconnects when trying
    to get to 13782] and the problem has been the client wasn't listening on
    13782 like it should be. The only workaround, this being a Windows client
    and no way to reset a port that I knew of, was to reboot. This fixed the
    problem.

    I suspect that the root cause was a piece of software [Altiris; used to do
    inventory and SW pushes] that was hogging up resources and not correctly
    releasing them and that led to problems with port connectivity.


    Scott

  5. Re: Cannot connect on socket


    Oh man... I'm having the same problem and I am stuck


    "Filbert Minj" wrote:
    >
    >Scott,
    >
    >I tried the steps you have mentioned from the client and I got the responses
    >like this:
    >
    >[root@scigate bin]# ./bpclntcmd -pn
    >expecting response from server ncsi-181
    >scigate.ncsi.iisc.ernet.in scigate.ncsi.iisc.ernet.in 10.180.1.14 33474
    >
    >[root@scigate bin]# ./bpclntcmd -hn ncsi-181
    >host ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    >checkhname: aliases:
    >
    >[root@scigate bin]# ./bpclntcmd -ip 10.180.13.15
    >checkhaddr: host : ncsi-181: ncsi-181 at 10.180.13.15 (0xf0db40a)
    >checkhaddr: aliases:
    >
    >Here ncsi-181 and 10.180.13.15 are hostname and ip of media server.
    >All these responses look ok.
    >
    >One more thing I noticed is when I tried to connect client using telnet


    >(telnet 10.180.1.14 13782), it closes the connection immediately where as
    >other clients do not. That means client is refusing the connection although
    >all the services (process) in clients are running.
    >
    >Do have any idea why this is happening?
    >
    >
    > wrote:
    >>
    >>"Filbert Minj" wrote:
    >>>
    >>>We have Veritas NetBackup 5.1 running on a window 2000 Server. Problem

    >started
    >>>when we migrated our servers to private IP address. It is not connecting
    >>>to a client, which is Redhat Linux 7.2 kernel 2.4.7-10. When I try to

    load
    >>>properties of this client from Host properties in the main console it

    gives
    >>>the error message "scigate.ncsi.iisc.ernet.in cannot connect on socket".
    >>>The status code is 25.
    >>>
    >>>We have several other Linux clients. One of these is Red hat Linux 7.1.

    >>It
    >>>works fine for all other clients.
    >>>
    >>>Any help will be appreciated.
    >>>Thanks
    >>>

    >>
    >>Filbert,
    >>
    >> The first thing I'd do is ping and traceroute to the Linux box from

    the
    >>master server and vice versa. If you can't do that, then NB isn't going
    >>to work.
    >>
    >> If that's successful, the problem could be that the client is no longer
    >>able to authenticate the NB server. This happened to me when I tried backing
    >>up some boxes that had been set up in a psuedo-DMZ: the DNS for the zone
    >>didn't have my NB servers in its database and therefore the authentication
    >>failed.
    >>
    >>From the client, cd /usr/openv/netbackup/bin and run
    >>
    >>bpclntcmd -pn
    >>
    >>bpclntcmd -hn media_server_name
    >>
    >>bpclntcmd -ip media_server_IP
    >>
    >>You should get responses back from each one verifying the master server

    >and
    >>the media server respectively.
    >>
    >>If this doesn't work, you might need to either configure a DNS in this

    address
    >>space or add host file entries to resolve the NB server names.
    >>
    >>
    >>Scott

    >



+ Reply to Thread