Frequent server synchronization errors?

Discussion in 'Windows Server Systems' started by ScottFern, Jul 12, 2007.

  1. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    We like a lot of networks share drives on our servers and our two servers DC2 and FS1 both have been giving users random synchronization errors. I usually have to open the synchronization utility and OK the synchronization and usually it synchronizes fine.

    By the way the Windows Firewall is turned off on all servers. This constant disconnects are affecting the users daily. I can't figure out what is causing this? I know before we had a Norton issue on DC2 and 70% of the errors are coming from that server. In fact we reinstalled AD on DC2 about 3 weeks ago.

    Norton causing this? Bad switch? I noticed one of the ports on one of the switches is bad.

    Now I did go around and imaged all workstations recently and did not RESET the computers in AD but it seems even after I do this it still comes up? HELP PLEASE!

    This is a Windows Server 2003 environment with XP workstations.
     
  2. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    Any chance this sounds like it is DNS related? We just decommissioned our ISA1 server today because it was acting up and are now on ISA2 and installed DNS on our DC2 server as well.

    I am getting sick of these Offline files need to synchronize messages constantly popping up all the time!
     
  3. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    I am beginning to think these offline files and synchronization errors are DNS related. Even our proxy server is having a hard time with name resolution. And I got this error this morning "Error Code: 502 Proxy Error. No data record is available. (11004)". Seems to me that periodically when Windows drive disconnect occurs and someone accessing their mapped drives and they reconnect that they are unable to find the server the drive is on thus creating a synchronization error as well as creating an offline files situation.

    I checked the DNS events and found nothing suspicious which is very frustrating for me but I need to find a lead here to check! Any help at all is appreciated.
     
  4. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    Ok, people new direction. How do I find name resolution errors when nothing is showing in event viewer of my DNS server? Could symantech cause these errors?

    My ISA server can't resolve the DHCP, Active Directory, or DNS servers as reported by the connectivity verifier!?
     
  5. Steevo

    Steevo Spammer representing. Political User Folding Team

    Messages:
    2,566
    Create a custom hosts file and use that, force IP addresses, or lengthen the lease time.




    It is the simplest way I have found. I setup my systems to have static IPs and just created my own hosts file. No more dropped Client Access Express sessions.
     
  6. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    Are you referring specifically to mapped drives? Because our servers that host files all have static IPs already. It seems to be more a situation of losing connection to those mapped drives.

    pardon my ignorance, but how does one change the hosts file? Are you referring to the LMHOSTS file?


    What if I cleared the DNS cache pool?
     
    Last edited: Jul 26, 2007
  7. Steevo

    Steevo Spammer representing. Political User Folding Team

    Messages:
    2,566
    Try forcing a loonger lease time for the clients, and or a static IP.

    Your clients hosts file, C:\Windows\System32\drivers\etc can be modified so there is no name lookup, it uses hosts over DNS. That will prevent DNS errors.


    Also look at client timeout in server management. I get that when guys don't sign out over the weekend and I try to perform maintainance. But since I don't force logoff I get file in use errors.
     
  8. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    You know I noticed that our main software program is JAVA based and the users when they reboot or log off are getting errors about synchronizing Qik.properties file and a HsspRegistry.ser file. Both files reside in the app data folders and they are redirected for ALL users.

    I am beginning to think that java doesn't understand the whole redirecting feature well?
     
  9. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    UPDATE ON THE ISSUE:

    I keep getting these two errors on a daily every other hour basis in ISA2004. I have checked the DNS event viewer and have found nothing alarming on hubdc1? What I have noticed is that every couple hours the Connectivity window on the Dashboard will report no connectivity for Active Directory, DHCP, DNS, and Others. I don't think its any surprise we host DNS and DHCP on hubdc1 and hubdc2 and get the error for all of them.

    Slow Connectivity
    Description: The connectivity verifier "DNS" reported a slow connection to hubdc1.hubtravel.dom. The specified connectivity response threshold was 5000 msec, and the monitored response time was 9110 msec.

    No Connectivity
    Description: The connectivity verifier "DNS" reported an error when trying to connect to hubdc1.hubtravel.dom.
    Reason: The request has timed out.

    I noticed also when this disconnect occurs we get a temporary 1-5 minute outage of the internet and I believe it to be DNS related. Any clues or hints where I should start to look?

    Thanks in advance!
     
    Last edited: Aug 6, 2007
  10. Steevo

    Steevo Spammer representing. Political User Folding Team

    Messages:
    2,566
    http://www.opendns.com/ (Thanks to Lord, the OSNN user, not well, you know)



    Try this, and on one machine use your hosts file for internal resolution.
     
  11. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    Turns out it was NOT dns issues. Rather it was a faulty switch and that was my original suspicion however my network admin kept saying prove it. Well I proved it when it locked up and all the switches couldn't transmit data........we now have 3 brand new switches! :)


    Thank god! I am happy my instincts were right.
     
  12. fitz

    fitz Just Floating Along Staff Member Political User Folding Team

    Messages:
    4,076
    Location:
    Chicagoland
    ? what kind of switches were you using?
     
  13. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    24 port Linksys SR244 and 4124. 2 each.

    Oh and I found out the hard way that you can't plug in two link cables from 1 switch another switch. Causes bad lockup. lol

    I already knew that but I didn't close enough attention to the cable I plugged in. Lesson learned.
     
  14. fitz

    fitz Just Floating Along Staff Member Political User Folding Team

    Messages:
    4,076
    Location:
    Chicagoland
    Depending on the switch, you can plug a straight-through cable to connect two switches - if the switches support auto MDI-X (looks like your 4124 should support that)

    Generally, if you are going to call anything really production, I would recommend a nice managed switch (HP Procurve, higher model Linksys even on the lower end, Cisco/Foundry/Nortel/Force10 on the higher end..)
     
  15. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    We are no longer using those switches. We are now running two Linksys SRW2048 and another Linksys 24 port Gigabit switch. All three are managed. What kind of configuration changes would be typical to make since its managed?
     
  16. fitz

    fitz Just Floating Along Staff Member Political User Folding Team

    Messages:
    4,076
    Location:
    Chicagoland
    not many for a smaller network. Might want to consider hardcoding ports to 100 full where necessary both on the switch and on the client side instead of relying on auto-negotation.

    If you want to get into the SNMP monitoring or set up a port mirror to monitor traffic, you could do that as well depending on the capabilities of the switch.
     
  17. Electronic Punk

    Electronic Punk Administrator Staff Member Political User Folding Team

    Messages:
    18,590
    Location:
    Copenhagen, Denmark
    Was going to suggest the switch.
    Fortunately we are incharge of both servers and switches so we don't have to fight as part of our diagnosis.
     
  18. ScottFern

    ScottFern IT guy

    Messages:
    48
    Location:
    Chicago
    Where were you over 2 months ago!? :p
     
  19. fitz

    fitz Just Floating Along Staff Member Political User Folding Team

    Messages:
    4,076
    Location:
    Chicagoland
    he was busy with his ex girlfriend traveling ;)