Changi NNTP Server 1.0

[ Bottom of Page | Previous Page | Next Page | Table of Contents ]

Recovering from System Crashes

Most probably the history file has been corrupted and the active file didn't get updated in case of a system crash. To recover from this condition run
expire -rvxs [add other parameters usually used on expiration]
after rebooting and before starting the server.

Systems running 24 hours a day may run this command on every reboot. Otherwise you may check for the existence of CHANGIPIDFILE, which will be removed by Changi during normal shutdowns. The existence of this PID file indicates that the server terminated abnormally.

Here is an example of NNTPD.CMD, which could be used to start Changi with a poper history index, even after system crashes:

/* Running Changi */
DO FOREVER
    '@expire -vVxs'
    IF rc \= 0 THEN
        '@expire -vrxs -E14'
    '@del changi.pid 2>NUL'
    '@changi'
    SAY 'Result=' rc
    IF rc = 0 THEN
        LEAVE
END


[ Top of Page | Previous Page | Next Page | Table of Contents ]


URL: changi/manual/probcrash.html
Created: 1 November 1996
Revised: 21 November 1996
Author: harald@os2point.ping.de