New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FB 2.0 RC1 BUGCHECK [CORE5] #323
Comments
Commented by: Alice F. Bird (firebirds) Date: 2006-05-04 18:43 The cause is the procedure request handling, it's currently |
Modified by: Dimitrios Chr. Ioannidis (dchri)Version: 2.0 RC1 [ 10031 ] assignee: Dmitry Yemanov [ dimitr ] SF_ID: 1481965 => |
Commented by: skaushanskiy (skaushanskiy) This problem happens in FB 1.5.3 Super Server on WIndows as well, STAN (Server) Thu Jun 15 09:15:42 2006 |
Commented by: @dyemanov Arno, any status update on this? |
Modified by: @dyemanovassignee: Dmitry Yemanov [ dimitr ] => Arno Brinkman [ arnobrinkman ] |
Commented by: skaushanskiy (skaushanskiy) The same problem also happens in FB 2.0 RC2, and FB 2.0 RC3. It happens when I am simultaneously inserting large numbers of records into the same tables from 70+ concurrent clients. I tried creating a test-case, but could not reproduce the issue. This issue has never happened to me under linux. My assumption is that it is a "threading" issue in the windows-server code. |
Modified by: skaushanskiy (skaushanskiy)Version: 2.0 RC3 [ 10040 ] Version: 2.0 RC2 [ 10032 ] Version: 1.5.3 [ 10028 ] SF_ID: 1481965 => |
Commented by: @pcisar Two questions: 1. Isn't this related to CORE538 ? |
Commented by: skaushanskiy (skaushanskiy) 1. No, in this issue, the FB server does not crash, it just throws BUGCHECKS and refuses requests for a few seconds, In CORE538 the FB server crashes but does not log anything, the guardian re-spawns the server, and FB acts like nothing happened. So yes, they are very similar, but the symtoms are a little different, I actually prefer CORE538 because FB does not drop client connections |
Commented by: skaushanskiy (skaushanskiy) About problems with PHP connectivity, This well known fact is not that well known, by me. can you give me more details about this? stan |
Commented by: skaushanskiy (skaushanskiy) Ok, I rewrote my code to use ibase.h and fbclient.dll directly in C/C++. The issue still occurs. I will work on a reproducible test case as time permits. So far my investigation suggests: It happens under heavy load (65+ concurrent long-running write transactions from different clients). It happends if I use the default transaction parameters and if I specify READ_COMMITED|NO_RECORD_VERSION. Again, this only happens under Windows. thanks, stan |
Commented by: skaushanskiy (skaushanskiy) Reproducible test case uses PHP 5.1.4 and Firebird 1.5.3. also happens under Firebird 2.0 RC 4. |
Modified by: skaushanskiy (skaushanskiy)Attachment: reproducible_BUGCHECK167.zip [ 10127 ] |
Commented by: skaushanskiy (skaushanskiy) I have a reproducible test case. It is attached. let me know if you need anything else. thanks, stan |
Commented by: skaushanskiy (skaushanskiy) This seems to be fixed in a "duplicate issue": this issue can be closed. |
Commented by: @pcisar It's a duplicate of fixed issue. |
Modified by: @pcisarstatus: Open [ 1 ] => Resolved [ 5 ] resolution: Fixed [ 1 ] Fix Version: 2.1.0 [ 10041 ] Fix Version: 2.0.1 [ 10090 ] SF_ID: 1481965 => |
Commented by: @pcisar Reopened to update ticket information. |
Modified by: @pcisarFix Version: 2.1 Alpha 1 [ 10150 ] SF_ID: 1481965 => Fix Version: 2.1.0 [ 10041 ] => |
Modified by: @pcisarWorkflow: jira [ 10029 ] => Firebird [ 14159 ] |
Modified by: @pavel-zotovQA Status: No test |
Submitted by: skaushanskiy (skaushanskiy)
Assigned to: @ArnoBrinkman
Is duplicated by CORE1080
Attachments:
reproducible_BUGCHECK167.zip
SFID: 1481965#
Submitted By: skaushanskiy
Configuration:
3GHZ P4 HT
1GB RAM
Windows XP SP2
My Database starts out small, but is always growing.
During the problem it was around 15MB
I had upto 60 simultaneous "local - php processes"
Clients connecting to the database and writing data
using stored procedures.
I am not doing any "create/drop" MDL statements,
only SELECTs, INSERTs, and UPDATEs
Under load I sometimes get:
PDD-STANK (Client) Tue May 02 13:40:29 2006
Guardian starting: C:\Firebird\bin\fbserver.exe
PDD-STANK (Server) Wed May 03 21:04:29 2006
Database: C:\DB.FDB
internal gds software consistency check (invalid SEND
request (167), file: exe.cpp line: 494)
This is not happening very often, so far happened
twice, in overnight testing, but I have only been
testing RC1 for 3 days.
I looked through the source and it looks like the
"request" pointer is NULL, but I have no idea why.
what could be causing this BUGCHECK?
Let me know if you need more information.
I am willing to try fixes.
thanks,
stan
The text was updated successfully, but these errors were encountered: