1. Get rid of all advertisements and get unlimited access to documents by upgrading to Premium Membership. Upgrade to Premium Now and also get a Premium Badge!

oracle 10.2.0.4 ssh/toad/sqlnet disconnects from remote access

Discussion in 'General' started by kwilkinson, Mar 26, 2010.

  1. kwilkinson

    kwilkinson Active Member

    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    55
    need assistance on this, users are remoting into company network using VPN client, VPN client maintains connectivity, but when users are using multiple instances of TOAD/SQLNet to access Oracle DB the idle connections timeout between 5-30 minutes. SSH also has same issue when accessing this way. problem occurred when we upgrade our firewall and switch. firewall router and switch have all been checked out, no log files of any errors from any of the hardware or on the server. oracle is running on Solaris Unix servers. users insist that this did not happen prior to upgrade hardware. we set the max timeouts as high as they would go on all hardware and still had same problem. Anyone have any ideas on this, please let me know. thank you.
     
  2. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    Check the following file on the database server

    $ORACLE_HOME/network/admin/sqlnet.ora

    for SQNET.EXPIRE_TIME. I expect that it is set and that setting may be causing the connection timeout issue. You should also query DBA_PROFILES for any profile with CONNECT_TIME set to a value other than UNLIMITED. These apply if the UNIX sessions are not timing out; if they are also suffering the same fate it is not an Oracle issue it's a server/router/switch issue or it's an issue with the ssh configuration.
     
  3. kwilkinson

    kwilkinson Active Member

    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    55
    Thank you for the info, i believe the Oracle DBA may have looked at these settings, he was able to come up with a temp resolution by having the Oracle db do a ping to the TOAD client that kept the sessions open.
    Are there any known problems with Oracle 10.2.0.4 and Juniper firewalls, or Ciso PoE(power over ethernet) 3750 switch?
     
  4. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    No. This isn't an Oracle problem since your ssh sessions are also affected. You should be checking the settings on the Cisco switch to see if you've missed some obscure configuration value causing this.
     
  5. kwilkinson

    kwilkinson Active Member

    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    55
    thank you for your reply, but we've been over the settings on the 3750 poE switch, the router, and the firewall. all timeouts on the hardware were set to the max setting and the problem is still occurring. The fact that they are using an older version of Oracle 10.2.0.4 and TOAD with new network hardware could not be an issue? also what about the MTU settings on the Oracle side, could that be an issue? Could an Solaris Unix update cause the issue with the new hardware? thank you for any info.
     
  6. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    MTU settings affect the packet transport not the timeouts. Again, this is NOT an an Oracle issue as Oracle has no control over ssh sessions which are also affected. Yes, a Solaris update could affect such behavior. You will need to look in places other than Oracle to solve this issue and the likely culprits are the O/S, switch, router and firewall.
     
  7. kwilkinson

    kwilkinson Active Member

    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    55
    Zargon,

    I appreciate your help on this, as i said we've been all over the switch, router and firewall, there is nothing in the log files anywhere indicating there's a problem, that includes on the Solaris server and the Oracle logs. let me re hash this, user's connect using VPN to our company network(this stays connected) then they use TOAD/SQLNet to access the Oracle DB, and the DBA uses the SSH. Since we can rule out the network switch, firewall and router after working on it for several months that leaves the Oracle/Solaris Unix server as the only other commonalities where the problem resides. This is why i asked about the MTU, i also have some experience with Unix servers running Oracle and have had issues when the unix patches are installed they move/remap the listener processes, this is where i think the problem lies. i am able to find plenty of blogs and sites like this where it is a known problem with TOAD/oracle dropping, but not with the same circumstances as I have. Thank you again for your assistance.