Shutdown Issue 6

[cbarrol]

Hello,
I have been having a problem lately when trying to down my Netware 6.0 server. (I restart it monthly during down time as part of a preventative maintenance schedule)

From the console prompt, after I type 'down' and acknowledge the shutdown message with a 'y', the system displays a message something like 'notifying clients that server is shutting down' then just hangs. I have to power it down manually with the on/off switch in order to restart it.

In the past this did not happen, the system would down itself gracefully. I have been fortunate in that it always comes back up, but I am looking for suggestions as far as what might be causing it to hang.

Thanks,
Craig Barroll

 

[terry712]

clear the connections before downing

cant understand why you are rebooting monthly - this is an alien concept - it's a netware server - only reboot it if replacing hardware or applyinmg latest sp.

 

[cbarrol]

Thanks for the quick reply

By clear the connections, I presume you mean via the monitor (F6 clear unused connections)

One reason for the reboot is to clear database locking files prior to reindexing our database. This must be done monthly. Often I get a sharing voilation if I try to reindex, even though all users are logged out, but I think this is a database issue rather than a Netware issue. the reboot clears the locked files.

 

[terry712]

CLEAR STATION ALL

this should do it as well

i would normally cron it

have a bounce.ncf

something like
CLEAR STATION ALL
delay 99
reset server

i think the delay comes from toolbox but not sure
i tend to cron through the browser now - as it's nice an easy fill in for - always crap getting correct syntax in the crontab

have things running at 2 in the afternoon instead of night etc

 

[marvhuffaker]

If you're on Novell's latest service pack, you can add the "-f" switch to the reset server command. this is supposed to force it down better than if you don't use that switch.

for example:

reset server -f
restart server -f
down -f




Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[terry712]

does the switch work ?

i want to cron a reboot of a few server this weekend as i notice the ip patch has changed - there is one or two that in the past tend to sit and smile at a cron#ed reset

if -f is good then i'll cron them all

 

[marvhuffaker]

I've had pretty good luck with them on 6.5. Not sure about 6.0, but 6.5 seems to hang more than I'm used to when it's down'ed and that switch has come in quite handy. It has worked every time I used it.

Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[allanl]

I had this problem with NetWare 6 and an Adaptec RAID controller. I don't think that there is a fix if you use some Adaptec's.

Try upgrading the Adaptec firmware?

 

[ITsmyfault]

fwiw delay is already there in sys\system.. at least that's where I found it. I use it in our Autoexec.ncf to make things load in an orderly fasion.
FWIW I agree with the others: rebooting is a Windows mentality. Don't accept it!! My Netware and Solaris servers are up for years in some cases without interruption (literally!!). If there are file locking issues, find out what's causing them - could be the DB software, could be a connections issue, or something else. But it's worth finding IMHO. What else could it affect? You just don't know until you know..
Due to the way the Netware kernel is architected it is easier to abend, strictly speaking, and as a result the code has to be written much cleaner than for a windows system. Not to say Provo is perfect (they're not), but rebooting is not part of "routine maintenance" and should not be. If anything, it tends to make netware problems worse by throwing sync farther out. (in the general case)
jm2c

on the downing issue, if nothing else works and it's making you crazy, you can send a coredump into Novell and have them figure out what's going on. Make sure your DIAG600 is loaded first. You can force an abend from the debugger and then take a coredump IIRC. There is a whole TID on it. Ad-hoc cases for folks w/o support is about $500 last time I checked. GD LK!

 

[BJJJB]

Locked Files requiring Maintenance.

Are you sure the DB fiels are flagged as Shareable, We've had problems with this and usually found someone has forgot to make them shared. We've found this improves network response times as well due to fewer resends being required.

I too have Netware Servers that are up for years, only taken down for upgrades, and thats rare. We've Adaptec RAID and despite disk failure we've managed to keep running.

BestOfLuck

BJJJB
"Murphy was an optimist"