• Article Type: General
• Product: Aleph
• Product Version: 16.02

Description:
The www_server is crashing. When we restart it, it stops again almost immediately. The log shows:

Execution error : file '/aleph/aleph/a16_1/aleph/exe/update_statistics.gnt'
error code: 114, pc=0, call=0, seg=0
114 Attempt to access item beyond bounds of memory (Signal 11)

(This message occurs several times.)

At the end of the log, when the www_server processes disappear from ps -ef and and from util w/1, we see:

server_main: OUT 0.0002 236
read 0 data from socket
read 0 data from socket
read 0 data from socket
read 0 data from socket
<etc.>

When the www_server is up, Internet Explorer users get a blank page and can do nothing; Firefox users can function normally except that the icons do not display. When the server is down, all users get "Page not found".

The problem started occurring after we rebooted the server.
What should be done?

Resolution:
This site has a separate server for Oracle (a "two-task" set-up). The problem occurred because the clustering software was pointing to the u50 tree binaries.

We saw that they had a running apache configured on the same port as Aleph 16 apache - port 83.

When they were killing the previous apache process, it was restarting itself (from the older version of the Apache).
This software was automatically starting the old Apache on port 83.

The server had not been rebooted since this change had been made (some months ago).

Once rebooted, the cluster checked the wrong apache bin directory and started a "rogue" apache process.

Note: This problem was entirely unrelated to the update_statistics program. (It's just that this is the first program which www_main calls....)