IIS-hosted ASP.NET app using Sybase DirectCONNECT intermittently hangs, requires app pool recycle

Question

IIS-hosted ASP.NET app using Sybase DirectCONNECT intermittently hangs, requires app pool recycle

Sridhar Kumar N 0

We are facing an intermittent production issue where an IIS-hosted application using the Sybase DirectCONNECT client becomes unresponsive (hung) once or twice every 1–2 weeks. The application does not crash, but all incoming requests remain stuck in “Processing” and never complete. Service is restored only after recycling the IIS application pool.

please refer the eventviewer screenshot, mutliple time this error comes - it auto resolves mostly but sometime hung for long and manaual APPPool Recycle needs to fix the issue.

AgaveJoe 31,266 Reputation points

2026-01-30T15:21:03.6833333+00:00

It sounds like you are dealing with a classic thread exhaustion or deadlocking issue, but there isn't quite enough information here to identify the root cause. To help narrow this down, could you provide the following details?

Environment Specifics:

Which version of the .NET Framework (or .NET Core/5+) is the application using?

What is the specific version of the Sybase DirectCONNECT client?

Are you using the OdbcConnection, OleDbConnection, or a specific Sybase managed provider (like AseConnection)?

Code Pattern:

Are you using async/await for your database calls? If so, are there any .Wait() or .Result calls that could be causing a context deadlock?

Can you confirm that all database connections, commands, and readers are wrapped in using blocks to ensure they are being disposed of and returned to the pool?

Diagnostics:

The Event Viewer screenshot you mentioned didn't seem to come through—could you describe the specific Error IDs or messages you see?

When the app is in a 'hung' state, have you checked the 'Worker Processes > View Current Requests' in IIS Manager? This will show exactly which URLs are stuck and for how long.

If possible, capture a memory dump of the w3wp.exe process while it is hung using Procdump or DebugDiag. This is usually the only way to see exactly where the threads are waiting (e.g., waiting on a socket read from Sybase or stuck in a lock).

Knowing if this is happening during a period of high traffic or after a specific database event would also be very helpful.
Sridhar Kumar N 0 Reputation points

2026-01-31T11:20:45.2566667+00:00
its only one URL -we are using to connect to mainframe and fetch data.

its not due to traffic.

DC is out of support now by SAP

.net framework 3.5 and 4.7 are used

out of below incident, only one incident stuck(1/27/2026 8:17:47 AM) and needed apppool restart...

Filtered: Log: Application; Levels: Critical, Error, Warning, Information, Verbose; Source: ; Event ID: 4786. Number of events: 64

Date and Time Source Event ID

1/30/2026 8:26:29 AM Complus 4786 (8)

1/29/2026 8:44:39 AM Complus 4786 (8)

1/28/2026 8:06:20 AM Complus 4786 (8)

1/27/2026 8:54:46 AM Complus 4786 (8)

1/27/2026 8:17:47 AM Complus 4786 (8)

1/26/2026 8:32:50 AM Complus 4786 (8)

1/26/2026 8:08:34 AM Complus 4786 (8)

1/26/2026 8:08:34 AM Complus 4786 (8)

1/26/2026 8:08:32 AM Complus 4786 (8)

1/26/2026 8:08:32 AM Complus 4786 (8)

1/23/2026 8:32:34 AM Complus 4786 (8)

1/22/2026 8:23:17 AM Complus 4786 (8)

1/21/2026 12:45:58 PM Complus 4786 (8)

1/21/2026 8:41:28 AM Complus 4786 (8)

1/20/2026 8:20:24 AM Complus 4786 (8)

ya, next time will caputure the dump and provide. can you analyse the dump , what kind of tool required for analysis?

1 answer

Your answer

AgaveJoe 31,266 Reputation points

2026-01-30T15:21:03.6833333+00:00

It sounds like you are dealing with a classic thread exhaustion or deadlocking issue, but there isn't quite enough information here to identify the root cause. To help narrow this down, could you provide the following details?

Environment Specifics:

Which version of the .NET Framework (or .NET Core/5+) is the application using?

What is the specific version of the Sybase DirectCONNECT client?

Are you using the OdbcConnection, OleDbConnection, or a specific Sybase managed provider (like AseConnection)?

Code Pattern:

Are you using async/await for your database calls? If so, are there any .Wait() or .Result calls that could be causing a context deadlock?

Can you confirm that all database connections, commands, and readers are wrapped in using blocks to ensure they are being disposed of and returned to the pool?

Diagnostics:

The Event Viewer screenshot you mentioned didn't seem to come through—could you describe the specific Error IDs or messages you see?

When the app is in a 'hung' state, have you checked the 'Worker Processes > View Current Requests' in IIS Manager? This will show exactly which URLs are stuck and for how long.

If possible, capture a memory dump of the w3wp.exe process while it is hung using Procdump or DebugDiag. This is usually the only way to see exactly where the threads are waiting (e.g., waiting on a socket read from Sybase or stuck in a lock).

Knowing if this is happening during a period of high traffic or after a specific database event would also be very helpful.
Sridhar Kumar N 0 Reputation points

2026-01-31T11:20:45.2566667+00:00

its only one URL -we are using to connect to mainframe and fetch data.

its not due to traffic.

DC is out of support now by SAP

.net framework 3.5 and 4.7 are used

out of below incident, only one incident stuck(1/27/2026 8:17:47 AM) and needed apppool restart...

Filtered: Log: Application; Levels: Critical, Error, Warning, Information, Verbose; Source: ; Event ID: 4786. Number of events: 64

Date and Time Source Event ID

1/30/2026 8:26:29 AM Complus 4786 (8)

1/29/2026 8:44:39 AM Complus 4786 (8)

1/28/2026 8:06:20 AM Complus 4786 (8)

1/27/2026 8:54:46 AM Complus 4786 (8)

1/27/2026 8:17:47 AM Complus 4786 (8)

1/26/2026 8:32:50 AM Complus 4786 (8)

1/26/2026 8:08:34 AM Complus 4786 (8)

1/26/2026 8:08:34 AM Complus 4786 (8)

1/26/2026 8:08:32 AM Complus 4786 (8)

1/26/2026 8:08:32 AM Complus 4786 (8)

1/23/2026 8:32:34 AM Complus 4786 (8)

1/22/2026 8:23:17 AM Complus 4786 (8)

1/21/2026 12:45:58 PM Complus 4786 (8)

1/21/2026 8:41:28 AM Complus 4786 (8)

1/20/2026 8:20:24 AM Complus 4786 (8)

ya, next time will caputure the dump and provide. can you analyse the dump , what kind of tool required for analysis?

Answer 1

Danny Nguyen (WICLOUD CORPORATION) 6,615 Microsoft External Staff Moderator

Hi,

This behavior typically indicates a hung/blocking Sybase DirectCONNECT call (often through COM+/Enterprise Services, consistent with Complus Event ID 4786) that never returns. That blocks ASP.NET request threads, so IIS shows requests stuck in Processing until an app pool recycle terminates the stuck threads and resets client/pooling state.

Key indicators from your details:

Only one URL triggers it → one specific DC code path is hanging.
Not traffic-related → not load, likely a stuck I/O/lock/COM call.
Repeated Event ID 4786 → COM+/Enterprise Services involvement is likely.

Next time it happens, please capture a dump of w3wp.exe before recycling:

DebugDiag (Hang rule) or ProcDump: procdump -ma <pid> C:\Dumps\w3wp_hang.dmp
Analyze with DebugDiag Analysis or WinDbg + SOS to see exactly what threads are waiting on (vendor DLL/COM wait/socket read/lock).

Mitigations:

Set explicit connection/login and command/query timeouts (avoid infinite waits).
Ensure all connections/commands/readers are always disposed (using) to prevent pool exhaustion.

If possible, please share the full text of one Event 4786 entry and confirm whether you’re using ODBC/OLE DB/managed provider/COM component.

Hope this helps.

Sridhar Kumar N 0 Reputation points

2026-02-02T14:17:46.59+00:00
Thanks Danny, please find the full logs,

Log Name: Application

Source: Microsoft-Windows-Complus

Date: 1/27/2026 8:17:47 AM

Event ID: 4786

Task Category: (8)

Level: Error

Keywords: Classic

User: N/A

Computer: cfe-prd-web-1.ad.services.abc.com

Description:

The system has called a custom component and that component has failed and generated an exception. This indicates a problem with the custom component. Notify the developer of this component that a failure has occurred and provide them with the information below.

Component Prog ID: 0[ODBC][Env 00C8C9C8]

Method Name: IDispenserDriver::CreateResource

Process Name: w3wp.exe

Exception: c0000005

Address: 0X77D964C7

Event Xml:

<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">

<System>

<Provider Name="Microsoft-Windows-Complus" Guid="{0f177893-4a9c-4709-b921-f432d67f43d5}" EventSourceName="COM+" /> <EventID Qualifiers="49154">4786</EventID> <Version>0</Version> <Level>2</Level> <Task>8</Task> <Opcode>0</Opcode> <Keywords>0x80000000000000</Keywords> <TimeCreated SystemTime="2026-01-27T14:17:47.997964200Z" /> <EventRecordID>102841</EventRecordID> <Correlation /> <Execution ProcessID="0" ThreadID="0" /> <Channel>Application</Channel> <Computer>cfe-prd-web-1.ad.services.abc.com/Computer> <Security />

</System>

<EventData>

<Data Name="param1">

Component Prog ID: 0[ODBC][Env 00C8C9C8]

Method Name: IDispenserDriver::CreateResource

Process Name: w3wp.exe

Exception: c0000005

Address: 0X77D964C7</Data>

</EventData>

</Event>
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-02-03T06:46:32.9133333+00:00
Thanks for the details.

Event ID 4786 is COM+ telling you that a custom/native component failed while COM+ was trying to create an ODBC resource:

Method: IDispenserDriver::CreateResource → COM+ Resource Dispenser creating/pooling an ODBC connection/resource

Prog ID: 0[ODBC][Env ...] → the failure is in the ODBC driver/DC layer, not typical managed (.NET) code

Exception: c0000005 → Access Violation (native memory fault). This usually points to a bug/corruption in the ODBC/DirectCONNECT driver (or one of its native DLLs), not something you can catch/handle reliably in .NET.

That matches your symptom: after this failure, new ODBC resources may stop being created correctly, requests pile up, and an app pool recycle clears the bad state.

Recommended: since DirectCONNECT is out of support, the most reliable fix is to upgrade/replace the DirectCONNECT/ODBC component (or isolate that one mainframe call into a separate process/app/service so it can’t stall the entire site).

Again, if this occurs again, the most useful confirmation is to capture evidence before recycling (e.g., a w3wp dump) so you can identify the exact faulting DLL for the 0x77D964C7 address and confirm it’s the DC/ODBC driver path.
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-02-18T10:35:57.6333333+00:00

Hi, just checking in, do you have any update on this post?
Sridhar Kumar N 0 Reputation points

2026-02-23T06:41:15.88+00:00

Many Thanks Danny,
Ya, we didn't get the hung yet, but we had similar latency issue. we took the dump & restarted the app pool where each transaction took 30-60seconds.

upon apppool restart it become milliseconds (100-300ms) response only :) .
anything from the dump can you put your expert analysis?
if so how to share the dump to you for analysis, its 482MB file.
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-02-24T03:09:30.51+00:00
Thanks for the follow up. I don't think it's advisable upload the dump directly. Instead, you can try to analyze it locally and paste the results here.

DebugDiag

Run DebugDiag Analysis (Hang/Crash analysis) on the dump.

Paste these sections from the report:

the blocked/waiting thread stacks

any stacks mentioning ODBC / DirectCONNECT / COM+ (resource dispenser)

the ODBC/DC-related DLL/module names + versions

WinDbg

Open the dump and paste a few representative thread stacks (sanitized), especially ones waiting in ODBC/DC code.

If this happens again, grabbing 2–3 dumps ~30 seconds apart before recycling can make the root cause much clearer.
Sridhar Kumar N 0 Reputation points

2026-02-25T13:46:52.57+00:00

Many Thanks for guidance. PFA.
Please reveiw and share your analysis. Thanks in advance/
MotoX80 37,516 Reputation points

2026-02-25T13:52:19.13+00:00

its only one URL -we are using to connect to mainframe and fetch data.

What does your database team see on the mainframe side at the point that it hangs?

If IIS sees 100 requests processing, does your DBA team see 100 queries active in the database? Do they have a tool like Toad for Oracle where they can analyze the active requests. Are connections doing full table scans or waiting on record locks or just connected?

Your problem might not be in IIS, it could be some other connection to the DB causing the requests coming from IIS to hang.

If they see that the connections from IIS are not doing anything, then that might indicate that the application code is not properly closing database connections, which in turn could lead to memory shortages in the app pool process.

Did this problem just start, or has it been an ongoing issue for months/years?
Sridhar Kumar N 0 Reputation points

2026-02-25T15:55:20.2+00:00

actually when we stop IIS pool and restart - issue disappared. so no mainframe issue clearly. it looks something queue in IIS /DC client side only seems
MotoX80 37,516 Reputation points

2026-02-25T16:18:05.26+00:00

so no mainframe issue clearly.

If you don't look to see what's running within the database instance, you won't know for sure.
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-02-27T10:50:28.1833333+00:00

Stopping/starting the IIS app pool definitely clears the symptom, but it doesn’t by itself prove the mainframe is not involved. Recycling simply kills w3wp.exe, drops/abandons connections, and resets COM+/ODBC/DC client state. So it can “fix” both a client-side stall and a server-side wait (locks/queued work) by forcing disconnect/retry.
Jack Dang (WICLOUD CORPORATION) 16,115 Reputation points Microsoft External Staff Moderator

2026-03-03T02:28:04.8966667+00:00

Hi @Sridhar Kumar N ,

I wonder if there is any update on this post. If you have any question, feel free to reach out.

If Danny's answer and his explanation were helpful - kindly follow the instructions here so others with the same problem can benefit as well.
Sridhar Kumar N 0 Reputation points

2026-03-03T12:18:16.6233333+00:00

Hi Jack Dang (WICLOUD CORPORATION) we are closely wathcing the issue and as its random issue waiting for hung to capture the dump. will update here once dump is collected for hung issue.
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-03-05T03:07:21.0266667+00:00

Thanks for the follow up. Let me know when you have any update! If you found our earlier discussion helpful or informative, I would greatly appreciate it if you could follow this guidance provide feedback. Thank you.
Sridhar Kumar N 0 Reputation points

2026-03-26T17:47:30.7833333+00:00

Danny, finally we have captured the dump trace as its 2 worker process, we captured both.
one is hung and other one was working that time in same IIS .

we have analysed with DebugDiag Analysis. and Good one is just 13 pages and hung one is 88pages.
unable to upload here as its over 3mb size.

is there a way to provide the report to you for analysis?
Danny Nguyen (WICLOUD CORPORATION) 6,615 Reputation points Microsoft External Staff Moderator

2026-03-27T02:34:37.94+00:00
Please copy/paste the key text sections here (sanitize hostnames/usernames/connection strings/data):

Summary page (top findings)

Thread Summary (counts of running/waiting threads)

Top 5–10 “stuck” threads with full call stacks (especially anything showing ODBC/DirectCONNECT/COM+/network waits)

Handle/Connection/COM sections (e.g., DB connection pooling / Resource Dispenser / number of open handles)

Any noted exceptions

If sharing the full report is still preferred, let me know and I will privately message you regarding this.

Share via

IIS-hosted ASP.NET app using Sybase DirectCONNECT intermittently hangs, requires app pool recycle

1 answer

Your answer