I'm having the same trouble and wondering if you have any other ideas.
I took your advice and am running the latest pg_foundry (2.0.0) ppm as well as latest Pg (8.3.3) and Activestate 5.8.x (522).
But I'm still having the issue. I don't have a better trace than the others, but I think I do know the type of behavior that is causing the problem. I've been running various load and stress tests, and it has only occurred when a lot of clients have killed the connections prematurely. The only evidence of this is in the pg log.
| [reply] |
ppm install http://pgfoundry.org/frs/download.php/1639/DBD-Pg-5.8.ppd
(or whatever version you want installed)
"same" problem? what problem? (The OP's "description" is far too opaque)
| [reply] [d/l] |
Event Type: Error
Event Source: Application Error
Event Category: (100)
Event ID: 1000
Date: 6/13/2008
Time: 4:46:48 AM
User: N/A
Computer: CAERUS
Description:
Faulting application httpd.exe, version 2.2.8.0, faulting module Pg.dl
+l, version 0.0.0.0, fault address 0x0000a9bb.
For more information, see Help and Support Center at http://go.microso
+ft.com/fwlink/events.asp.
Data:
0000: 6c707041 74616369 206e6f69 6c696146
0010: 20657275 74746820 652e6470 32206578
0020: 382e322e 6920302e 6750206e 6c6c642e
0030: 302e3020 302e302e 20746120 7366666f
0040: 30207465 61303030 626239
This happens intermittently during load testing, and I think I've traced to what type of load testing does it. Using pylot or ab, it doesn't crash. But, if I do the following procedure, I can get it to crash:
- Save a bunch of tabs in a browser, in my case I'm using Safari for Windows 3.1.1 (525.17).
- Click on the bookmark of the tabs over and over, before they get done rendering.
- I think safari is canceling the client requests or somehow dropping them mid-way through. I presume this is a factor.
| [reply] [d/l] |
Thank you! I worked a few hours on this - your advice worked perfectly. | [reply] |