Questions about Informix Turbo

We have experienced some problems with Informix Turbo 2.10.03G: Please excuse me if this has already been throughly discussed. Maybe I wasn't listening real good because I thought it didn't effect us. We have been trying to back-up our logs at night out of cron using a shell script program. SHELL SCRIPT run by root in cron from this line: ulimit 40240000; su informix -c "/usr/informix/local/bkup" shell script: cp /dev/null /bkup1/dblogs/dblog SQLEXEC=/usr/informix/lib/sqlturbo export SQLEXEC /usr/informix/bin/tbmode -u /usr/informix/bin/tbtape -a > /tmp/bkup.out <<+ + mv /bkup1/dblogs/dblog /bkup1/dblogs/log.ar cp /dev/null /bkup1/dblogs/dblog chmod 666 /bkup1/dblogs/dblog /usr/informix/bin/tbtape -s >> /bmp/bkup.out <<+ + mv /bkup1/dblogs/dblog /bkup1/dblogs/archive cp /dev/null /bkup1/dblogs/dblog chmod 666 /bkup1/dblogs/dblog /usr/informix/bin/tbmode -m /usr/informix/bin/tbstat -d >> /tmp/bkup.out Additional info: Our primary "Turbo Trained" system administrator has been off this week and we know very little about Turbo Informix. We are not really sure the Logs are writing to disk but the Archive seems to be working when it doesn't run out of disk space, which it did last night. Occasionally it runs out of disk while doing the Archive and blows off. The tape device that we have set is /bkup1/dblogs/dblog. We have had the block size set to 1024. We were having a read error 14 when we tried to do a log backup from within tbmonitor or manually executing the commands in the script file above. We could not back up the logs. We took the Unix system down and back up again several times. We also tried to change to the 9-track tape drive as the device and back up to tape. Nothing was successful. We finally contacted someone at another Army installation and they told us to change the block size in the tape parameter to 32 and the log back-up would work. It did!! but we decided to also Archive and it blew out while attempting the archive. The person we contacted mentioned that they had run Turbo just fine for several months with block size 1024 and then all at once they had to change their block size to 32 and that they continue to run with the block size at 32. Then... We tried to do an "tbinit -s" after the Archive blew out and had to kill it on the terminal because it never came back. We could not even do an "ls -al" on /bkup1. That "ls -al" became an unkillable process and we had to take the system down. We took the Unix system down and brought it back up and the database came up in Turbo just fine. Our customer is running now. However we need to do an Archive and wonder why all this has happened. After the system wouldn't tbinit with 32 as block size we changed the block size back like it was to 1024. We are "RUNNING" so guess that is ok. Could we have left the block size at 32? What was probably causing us to not be able to back-up? Finally, please explain what is happening when the logs are being backed-up? Is it necessary to do both an archive and a log back-up everyday? Does an archive do both, back the database up and clear the logs or do the logs have to be backed up first and then do an archive? Any comments would be greatly appreciated. Thanks in advance for your patience. Cheryl Gross cheryl@redriverad-emh1.army.mil (903) 334-3518 (We are government and on a very tight budget - we don't decide what to buy, we don't write the programs (sometimes) and we don't get the credit when it works, just when it doesn't) (Above is the opinion of ... and not that of my employers)