medv
asked on
SBS 2003 R2 stops responding to clients
Urgently looking for help on a SBS 2003 R2 Premium Server which suddenly stops to respond to clients.
After 10 months of normal operation the server shows the following
The symptoms start independent of client access or network load, sometimes the problem starts with all clients shut down.
The Server which is the only server in the network(DC, DNS, DHCP, Exchange, SQL2005) running a sql 2005 database suddenly stops responding to client access. SQL Server is not reachable and client logon fails.
The chronology is the same all 1 or 2 days and arises at random times.
The eventlog first shows an MSExchangeDSAccess Event ID 2102 Error in the application log meaning DC Server not responding
after aprox. 18 seconds it shows an MSExchangeDSAccess Event ID 2103 (GC- Server not responding)
These event ID´s are followed by MSExchangeAC 8026 Error(Server is shut down) and MSExchangeSA Error 9154(Error Code 0x80004005)
Then there are multiple Exchange errors
aprox. 15 Minutes after the first Error in the Application Log the AD Log shows a NTDS General Event ID 1655 (AD connection to GC failed) following by a NTDS General Event ID 1126(AD Connection to GC failed Error: 1460 timeout)
1 hour later there is an event entry 1869(AD found GC at "ServerName") followed by Event ID 1655 and 1126 again.
aprox. 1 minute after the error occurs the DNS log shows an Event ID 4015 (critical Error in AD) followed by 4x Event ID 4004 and then many Event ID´s 4000(DNS server could not open AD)
The System log shows no Errors.
The only solution to get the server back to work ist to do a restart, but the server doesn´t shut down completely for hours(hanging at blank screen with blue background(no BSOD)
After a hard reset the server comes up normally and works without problems for one or two days, till the same symptoms appear again.
System:
SBS 2003 R2 SP2 all updates applied
DCDiag and NetDiag don´t show any errors, Chkdsk is running without errors on all harddrives
Hardware:
Intel DQ35JOE with Core2Duo E6600
onboard VGA
3x Seagate SATA HDD on onboard RAID
Desperately looking for further help
thomas hajek - austria
After 10 months of normal operation the server shows the following
The symptoms start independent of client access or network load, sometimes the problem starts with all clients shut down.
The Server which is the only server in the network(DC, DNS, DHCP, Exchange, SQL2005) running a sql 2005 database suddenly stops responding to client access. SQL Server is not reachable and client logon fails.
The chronology is the same all 1 or 2 days and arises at random times.
The eventlog first shows an MSExchangeDSAccess Event ID 2102 Error in the application log meaning DC Server not responding
after aprox. 18 seconds it shows an MSExchangeDSAccess Event ID 2103 (GC- Server not responding)
These event ID´s are followed by MSExchangeAC 8026 Error(Server is shut down) and MSExchangeSA Error 9154(Error Code 0x80004005)
Then there are multiple Exchange errors
aprox. 15 Minutes after the first Error in the Application Log the AD Log shows a NTDS General Event ID 1655 (AD connection to GC failed) following by a NTDS General Event ID 1126(AD Connection to GC failed Error: 1460 timeout)
1 hour later there is an event entry 1869(AD found GC at "ServerName") followed by Event ID 1655 and 1126 again.
aprox. 1 minute after the error occurs the DNS log shows an Event ID 4015 (critical Error in AD) followed by 4x Event ID 4004 and then many Event ID´s 4000(DNS server could not open AD)
The System log shows no Errors.
The only solution to get the server back to work ist to do a restart, but the server doesn´t shut down completely for hours(hanging at blank screen with blue background(no BSOD)
After a hard reset the server comes up normally and works without problems for one or two days, till the same symptoms appear again.
System:
SBS 2003 R2 SP2 all updates applied
DCDiag and NetDiag don´t show any errors, Chkdsk is running without errors on all harddrives
Hardware:
Intel DQ35JOE with Core2Duo E6600
onboard VGA
3x Seagate SATA HDD on onboard RAID
Desperately looking for further help
thomas hajek - austria
Have you checked the hardware? do a chkdsk, defrag, and run system diagnostics for the vendor.
ASKER
@RobWill:
The problem occured long time after installing sp2. None of the symptoms mentioned in the KB93659 article are there. I don´t think that the problem is tcp or network related, because the server is reachable over the network(RDP, remote administration via Radmin V3, ping, telnet...). It seems to be an os internal problem where the server looses connection to its own GC and then the dc role doesn´t work.
@loftyworm:
All hardwaretests(memory, harddisks, chkdsk,...)showed no errors. No BSOD or instable behaviour.
The problem occured long time after installing sp2. None of the symptoms mentioned in the KB93659 article are there. I don´t think that the problem is tcp or network related, because the server is reachable over the network(RDP, remote administration via Radmin V3, ping, telnet...). It seems to be an os internal problem where the server looses connection to its own GC and then the dc role doesn´t work.
@loftyworm:
All hardwaretests(memory, harddisks, chkdsk,...)showed no errors. No BSOD or instable behaviour.
I ran across this a while back, it may help you;
Trouble shooting a hung server;
http://searchwincomputing.techtarget.com/tip/0,289483,sid68_gci1312855,00.html?track=NL-1005&ad=639465&asrc=EM_NLT_3632059&uid=787865#
Trouble shooting a hung server;
http://searchwincomputing.techtarget.com/tip/0,289483,sid68_gci1312855,00.html?track=NL-1005&ad=639465&asrc=EM_NLT_3632059&uid=787865#
ASKER
@loftyworm:
the only time the server really seems to hang is when i try to restart when my problem occured. I´m not sure to be able to initiate a crashdump at this time because it seems to hang at the last actions of the shutdown(settings saved, network connections stopped and gui already terminated). Since this server is in daily use and the one and only server in the network i mostly have to do remote administration. My customer will go for holidays in 2 weeks and then i can take this machine to my office and do further tests.
But oncde again: it doesn´t hang, it only looses some functionality
the only time the server really seems to hang is when i try to restart when my problem occured. I´m not sure to be able to initiate a crashdump at this time because it seems to hang at the last actions of the shutdown(settings saved, network connections stopped and gui already terminated). Since this server is in daily use and the one and only server in the network i mostly have to do remote administration. My customer will go for holidays in 2 weeks and then i can take this machine to my office and do further tests.
But oncde again: it doesn´t hang, it only looses some functionality
I just noticed the mother board you spec'd, the DQ35JO. You are aware this board does not support server O/S's, and there are no drivers for Server O/S's, are you? this could be related to the problem.
Also Intel recommends with that board if having network issues the following update be applied if it hasn't been already.
http://www.microsoft.com/downloads/details.aspx?FamilyID=17d997d2-5034-4bbb-b74d-ad8430a1f7c8&DisplayLang=en
Also Intel recommends with that board if having network issues the following update be applied if it hasn't been already.
http://www.microsoft.com/downloads/details.aspx?FamilyID=17d997d2-5034-4bbb-b74d-ad8430a1f7c8&DisplayLang=en
ASKER
@RobWill:
I know Q35 Chipset Boards not to be recommended for Windows Server 2003, but we selected this board because of its remote management capabilities(VPro technology) and because of the lower costs using desktop hardware.
I built 7 Servers identical to the erratic machine with same hardware, os and identical patch level for different customers. All working well exept this one. So i don´t think the problems are mainboard related(but who knows).
The KB884020 fix will not apply because its designed for WinXPSP2 not for Server 2003. I don´t think the problems are network related because remote access to the server works all the time, it seems the server itself looses the ability to connect to the AD Database or the GC.
I know Q35 Chipset Boards not to be recommended for Windows Server 2003, but we selected this board because of its remote management capabilities(VPro technology) and because of the lower costs using desktop hardware.
I built 7 Servers identical to the erratic machine with same hardware, os and identical patch level for different customers. All working well exept this one. So i don´t think the problems are mainboard related(but who knows).
The KB884020 fix will not apply because its designed for WinXPSP2 not for Server 2003. I don´t think the problems are network related because remote access to the server works all the time, it seems the server itself looses the ability to connect to the AD Database or the GC.
You are building and selling desktop board servers to clients? That is definitely not recommended Not meaning to lecture, but that could come back to bite you when a client finds out their configuration is not supported by the manufacturer. A test environment is one thing, but production environment is risky. It's conceivable a BIOS upgrade could bring down the server and their would be no support by Intel.
ASKER
@RobWill:
I agree with you, but one of the reasons to do so were the much lower costs and the needed VPro technology. Of course my customer was informed about the possible risks to do so. Beside the fact we didn´t use a dedicated and supported Server board the system was up for over 7 months without a problem and 7 identical "Servers" are doing very well.
For my customer it was unacceptable to build a comparable solution with dedicated server components.
No it has be done and i wil have to try to find a solution.
Further hints to diagnose my problem are welcome
Best Regards
Thomas
I agree with you, but one of the reasons to do so were the much lower costs and the needed VPro technology. Of course my customer was informed about the possible risks to do so. Beside the fact we didn´t use a dedicated and supported Server board the system was up for over 7 months without a problem and 7 identical "Servers" are doing very well.
For my customer it was unacceptable to build a comparable solution with dedicated server components.
No it has be done and i wil have to try to find a solution.
Further hints to diagnose my problem are welcome
Best Regards
Thomas
So long as the client is aware, I don't see it as a problem. Though as mentioned it is possible a BIOS upgrade could make one of your XP drivers, fail under 2003.
Any chance the server has multiple NIC's or multiple addresses assigned to 1 NIC?
Ps- If looking for good deals on SBS servers, HP often has some "best buy" models that are only available through their dealers. Two of my clients just bought very basic units for 5 user networks, and I was blown away by the price. They included Intel server motherboard, single dual-core Xeon 2.66 processor, 4 GB ECC RAM, 2 160GB SATA drives, SBS 2003 w/ 5 CALs for $1299 each. HP, Microsoft, and I believe IBM package these entry level systems every June. I don't know if it continues beyond that. Basically you get SBS for free.
Any chance the server has multiple NIC's or multiple addresses assigned to 1 NIC?
Ps- If looking for good deals on SBS servers, HP often has some "best buy" models that are only available through their dealers. Two of my clients just bought very basic units for 5 user networks, and I was blown away by the price. They included Intel server motherboard, single dual-core Xeon 2.66 processor, 4 GB ECC RAM, 2 160GB SATA drives, SBS 2003 w/ 5 CALs for $1299 each. HP, Microsoft, and I believe IBM package these entry level systems every June. I don't know if it continues beyond that. Basically you get SBS for free.
ASKER
Only one NIC and only one IP.
Changing the hardware would be the last option in my case.
Found similar behaviour in a posting of MS SBS Newsgroup on another server(loosing AD and GC connection and DNS functionality)
not sure if the problems are identical but trying to investigate and compare the setups.
Changing the hardware would be the last option in my case.
Found similar behaviour in a posting of MS SBS Newsgroup on another server(loosing AD and GC connection and DNS functionality)
not sure if the problems are identical but trying to investigate and compare the setups.
>>"Changing the hardware would be the last option in my case."
Sorry, wasn't suggesting that, just something to keep in mind with future systems. I was impressed with the pricing as an alternative for small "workgroup" offices.
Sorry, wasn't suggesting that, just something to keep in mind with future systems. I was impressed with the pricing as an alternative for small "workgroup" offices.
ASKER
I installed all MS updates released last week and installed the newest Intel matrix storage manager(using ich9r onboard raid) . Now 3 days without any misbehaviour.
Excellent. Good to hear.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Thanks for posting your findings medv. Glad to hear you were able to make some improvements.
Cheers !
--Rob
Cheers !
--Rob
I noticed you have SP2 installed. Numerous network related problems have been documented with SP2 on SBS, mostly related to it's use of Receive Side Scaling. The following article outlines the problems and has a link to a patch, which may be of some help:
http://support.microsoft.com/kb/936594
You may want to review other common issues with SP2
http://www.smallbizserver.net/Default.aspx?tabid=53&forumid=6&postid=67633&view=topic