Urgent! ds.nlm missing and server wont boot 1

[benbeneke]

Hi,

I have a single standalone 4.11 server running Groupwise 5.2 as well.

After some GW database problems today I got an abend in GWENN1.nlm which points to a hardware failure on NW's support site.

Now when I down and restart the server it will not complete startup! It asks for the server name and IPX Net name - HELP!

I can get the server through stage 1 (pre startup.ncf) and startup.ncf but stage 2 (pre autoexec.ncf) loads fail.

It all points to SYS:System directory not being there as a lot of load commands wont work.

Any ideas as I need this server back up in 2 days (thank god for the weekend).

I am currently trying to get a cdrom to load in dos so as to try reinstalling 4.11 enough to get the missing files back. Good or bad idea??

Any help appreciatted.

Ben

PS Server is very old:
PP200 with 128MB ram, 2 x 4.5GB SCSI DUplexed drives
Due for replacement and upgrade after Xmas, just my luck

 

[igortsin]

It looks like server cannot mount volume SYS:
Try run vrepair and mount volume.

 

[marvhuffaker]

Hard Drive failures suck, especially when you only have one server in the tree. not only do you lose the data itself, but all of your tree info.

Hopefully a vrepair will fix it, but if not -- you are in a tight spot. Hope you at least have a good and recent backup. Maybe if you're lucky the duplexed drive will be have your data and it won't be as painful.

BTW, reinstalling won't help you out here.. As terry mentioned, the SYS volume is not mounting. If you could get it to mount, your files will be there. If the volume is hosed, reinstalling Netware will not fix it. (That might be a myth started by pro-MS folks)



Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[benbeneke]

Hi Guys,

Well Xmas interrupted play but thought I would pass on what I did so as to help others.

This apparently cannot be fixed but I did it with some intuition, luck and a good pointer from the NW knowledgebase (TID10018811).

SYS\System directory had vanished (corruption?? - who knows!) So therefore none of the nlm's where there to load.

Thanks god for the up to date print outs I had of the following in my server ifo folder:

Startup.ncf
Autoexec.ncf
netinfo.cfg

First thing was to restore the system folder to a different server (no replica ring as I have 1 4.11 and 1 6.0 server)

Copy all .nlm files from the restored system folder to DOS C: drive on the affected machine.
I did it the very slow and hard way via floppies (no cd drive)

at c: prompt type:

server -ns -na which starts the server without startup.ncf or autoexec.ncf

Now following the above mentioned TID and your printouts of the Startup.ncf, Autoexec.ncf and netinfo.cfg start loading the nlm's manually.

This is long and tiresome as some nlm's must be loaded before others in the ncf's but are not listed as the server normally takes care of it in the background so I had a lot of going back to the beginning and starting again. once you get to the netinfo.cfg you are winning!

Load and bind the network card to TCPIP and more importantly (for me) IPX

You should now be able to see the server from an external PC. At this point I copied all my mailbox files off the server mail_vol to cover myself then copied the restored system file into sys on the affected machine.

Now down the server and restart - Everything came back up as normal!!!!

I ran dsrepair to make sure of any faults and ordered a replacement INTEL whitebox on the Monday morning.

The whole episode was painful but still satisfying to find a fix where none was supposed to exist.

Hopefully this may help someone else in the future.

Ben

 

[karmic]

Have a star for that one. Good one for future useage.

~ K.I.S.S - Don't make it any more complex than it has to be ~