Database disconnects (SQL + Oracle on VM)

Hi list, we recently moved 100% to VM for our three instances (2 instances JDE ERP8 SP24.1.x, 1 instance 8.12 TR8.98 / SQL2008 x64 + Oracle 11g, Win2008 x64). We receive on all three instances (from time to time) some DB disconnects. And as you know this is hard to trouble shoot (investigation ongoing etc.). I was wondering if anyone is using VM (on similar technology stack) and expirence same
issues on JDE. Thanks in advance, cheers Adrian

Hi, I'm getting the same thing. My database server is a physical SQL server, and my application (or enterprise server) is virtual. All Windows. We're on JDE 9.0. It doesn't happen very often, but I've seen it 3 or 4 times over the past few months. And it only seems to cause problems for the scheduler kernel. The other kernels show the same problem (database disconnect) in the log file, but seem to recover on their own. But after the error occurs, our scheduler sometimes errors out when trying to launch a job, but not always. So it will fail to launch one UBE, then a few hours later it will successfully launch one, then fail again. Strange indeed. My only resolution is to manually kill the scheduler process using the PID on the windows server, then restart it in JDE.

Here's an excerpt from one of my log files:

2344/2452MAIN_THREAD Wed Sep 21 16:02:03.014000Jdbodbc.c8437
ODB0000163 - wSQLExecute failure. rc=-1
2344/2452MAIN_THREAD Wed Sep 21 16:02:03.045000Jdbodbc.c8437
ODB0000164 - STMT:00 [08S01][10054][2] [Microsoft][SQL Server Native Client 10.0]TCP Provider: An existing connection was forcibly closed by the remote host.
2344/2452MAIN_THREAD Wed Sep 21 16:02:03.060000Jdbodbc.c8437
ODB0000164 - STMT:01 [08S01][10054][2] [Microsoft][SQL Server Native Client 10.0]Communication link failure
2344/2452MAIN_THREAD Wed Sep 21 16:02:03.076000Jdb_drvm.c1128
JDB9900401 - Failed to execute db request
2344/2452MAIN_THREAD Wed Sep 21 16:02:03.076001Jtp_cm.c1347
JDB9900255 - Database connection to F91300 (System - 900) has been lost.
2344/2452MAIN_THREAD Wed Sep 21 16:02:03.139000Jtp_cm.c1301
JDB9900256 - Database connection to (System - 900) has been re-established.

We've experienced the same thing. Our configuration was Microsoft Hyper-V utilizing 10 gig NICs connecting to a Cisco Nexus switch. Although we've not completely eliminated the problem we have greatly mitigated it through the following actions:

1.) We upgraded all hosts to the latest version of the network driver and firmware.

2.) Within the driver on each host we doubled the size of both the RX and TX buffer. What was occurring is that the buffer woukld almost fill and then the NIC driver within the switch port an RXDELAY asking not to send any more data until it could process it with buffer.

3.) We upgraded the version of IOS on the Nexus switches. The version we were running had issues with how it handled flow control.

These three items reduced our occurrences of the problem by about 90%.

Hi Paul,

Thanks for the tips. What you did makes sense, and I'll have to take at look at those things. I did some more digging yesterday afternoon, and discovered that the VM was migrated at exactly the time when I see all the error logs in the JDE log files on the enterprise server. So I have a feeling that could be the culprit. I'll let you know if I see it again, but it could be weeks before it happens again.

You will see a momentary disconnect during a migration.

Andrew,

Did you able to resolve this issue. I am having similar issue

E9.0,
SQL Server 2008 R2, Windows 2008 R2 (64 Bit) - Physical Server.

Enterprise Server , Windows 2008 R2 - Virtual Server.

Regards,