Subject | IB/Firebird server abnormal termination, server and client 10054 errors |
---|---|
Author | tomasz007 |
Post date | 2003-06-13T23:46:50Z |
Hi all,
Before writting this post I've gone through similar threads in the
IB* newsgroups, but haven't found the clear explanation/solution.
configuration:
--------------
- server: Win2K server + WI-V6.2.908 Firebird 1.0
- DB file: around 20meg, backed up everyday but not restored, gfix
does not report any errors, no UDFs installed
- 10+ concurrent clients: Win2K Pro + WI-V6.2.908 Firebird 1.0 + BDE
5.1.1 + application developed in D5
- network: some 50 win2K/XP workstations, MS Network, mostly only
TCP/IP bound to NICs, fixed IP addresses
NOTE: previously there was IB 6.02 opensource installed and the
problem started to happen and I changed it to Firebird hoping this
would help...
NOTE2: one strange thing with the DB is that if I do backup and
restore I have to drop and recreate one of the views, because
selecting from it gives "arithmetic....truncation" error (I use
WIN1250 collation)
NOTE3: I cannot reproduce the problem in my oficce network
NOTE4: The application has worked at customer for last 2,5 years fine.
During last 6 months a lot of additional functionality has been added,
customer connected more workstations to the network and more of them
uses the application
NOTE5: that the sales automation system - a lot of DB traffic is being
generated by the application (up to 2MB by one client)
problem:
--------
since some time ago, from time to time, with (of course!) different
frequency in different application places (operations) the application
gets closed with message that it generated and error and will be
closed by windows. At the same time all other client apps on other
machines behave exactly in the same way. The interbase.log from the
server shows:
SYSMAG-SERVER (Client) Mon Jun 09 15:06:27 2003
C:\Program Files\Firebird\bin\ibserver.exe: terminated abnormally
(-1)
in both - clients' and server's interbase.log files there are many
10054 errors, but not exaclty at the same time when abnormal
termination happens - i.e:
SYSMAG-SERVER (Server) Mon Jun 09 15:23:19 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:23:19 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:41:32 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:41:32 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 16:12:01 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 16:12:01 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Client) Mon Jun 09 17:50:06 2003
C:\Program Files\Firebird\bin\ibserver.exe: terminated abnormally
(-1)
my application's log file does not indicate any particular operation
that could trigger this problem - as I wrote before.
Recently I've even implemented some 'smart' way the app handles the
server restart (tested by manual server restart) and displays some
polite message, but this does not happen (of course!) at the customer.
As I wrote - the app gets kicked out by Windows.
Questions:
----------
- what may be the cause
- if network problem - whats the best way to trace it
- have somebody had (and solved) similar problem
The customer gets fed up - please help :)
Cheers,
Tomasz
Before writting this post I've gone through similar threads in the
IB* newsgroups, but haven't found the clear explanation/solution.
configuration:
--------------
- server: Win2K server + WI-V6.2.908 Firebird 1.0
- DB file: around 20meg, backed up everyday but not restored, gfix
does not report any errors, no UDFs installed
- 10+ concurrent clients: Win2K Pro + WI-V6.2.908 Firebird 1.0 + BDE
5.1.1 + application developed in D5
- network: some 50 win2K/XP workstations, MS Network, mostly only
TCP/IP bound to NICs, fixed IP addresses
NOTE: previously there was IB 6.02 opensource installed and the
problem started to happen and I changed it to Firebird hoping this
would help...
NOTE2: one strange thing with the DB is that if I do backup and
restore I have to drop and recreate one of the views, because
selecting from it gives "arithmetic....truncation" error (I use
WIN1250 collation)
NOTE3: I cannot reproduce the problem in my oficce network
NOTE4: The application has worked at customer for last 2,5 years fine.
During last 6 months a lot of additional functionality has been added,
customer connected more workstations to the network and more of them
uses the application
NOTE5: that the sales automation system - a lot of DB traffic is being
generated by the application (up to 2MB by one client)
problem:
--------
since some time ago, from time to time, with (of course!) different
frequency in different application places (operations) the application
gets closed with message that it generated and error and will be
closed by windows. At the same time all other client apps on other
machines behave exactly in the same way. The interbase.log from the
server shows:
SYSMAG-SERVER (Client) Mon Jun 09 15:06:27 2003
C:\Program Files\Firebird\bin\ibserver.exe: terminated abnormally
(-1)
in both - clients' and server's interbase.log files there are many
10054 errors, but not exaclty at the same time when abnormal
termination happens - i.e:
SYSMAG-SERVER (Server) Mon Jun 09 15:23:19 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:23:19 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:41:32 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 15:41:32 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 16:12:01 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Server) Mon Jun 09 16:12:01 2003
INET/inet_error: read errno = 10054
SYSMAG-SERVER (Client) Mon Jun 09 17:50:06 2003
C:\Program Files\Firebird\bin\ibserver.exe: terminated abnormally
(-1)
my application's log file does not indicate any particular operation
that could trigger this problem - as I wrote before.
Recently I've even implemented some 'smart' way the app handles the
server restart (tested by manual server restart) and displays some
polite message, but this does not happen (of course!) at the customer.
As I wrote - the app gets kicked out by Windows.
Questions:
----------
- what may be the cause
- if network problem - whats the best way to trace it
- have somebody had (and solved) similar problem
The customer gets fed up - please help :)
Cheers,
Tomasz