Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations TouchToneTommy on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

NDS Sync / Replica problem.

Status
Not open for further replies.

Datong

IS-IT--Management
Feb 15, 2005
118
GB
Im working on a small network of 17 NW6 (sp5)servers in a tree. The WAN links are slow and a bit dodgy. Looking at replica sync in dstrace there are errors popping up reporting on a few servers saying that the root parition is not on. A quick look at the servers themsleves and nds manager clearly shows that they are there and on. (no errors). One site has a replica stuck in the new state and every attempt to resolve it has failed. Dstrace shows errors 603, 654, 673, 672 at variou stages but none of the fixes seem to be working. Any ideas ?
 
have you just put the replica on?

stuck new states are a pain but dependant on link state should resolve - but even 64k should be reasonable - as it is a small tree ?
 
Holy crap. You have serious issues.

NDS Manager is not reliable in this case because it's only looking at one replica. You need to see the status from all servers, which you do at the server level.

Just as a word of caution, you are in seriously bad shape and you could completely destroy your NDS if you do the wrong things here. You need help. Unfortunately Novell does not support NetWare 6.0 anymore.

654 errors mean that a partition operation never completed. The root partition wont' come on till this error has cleared. There are other reasons why the operation never finished, so the 654 is only a symptom of a much deeper problem.

DSTRACE doesn't fix anything, it just shows you exactly what DS is doing. Those errors are insane. You need professional help.

My company can help you, but don't think that I'm trying to solicit your business. It is just that it sounds like you are in over your head here. A lot of customers let it go too long, and damage the system severely before they call us. If they had called us first, they could avoid this and avoided a lot of pain and agony. Sometimes it costs 2-3 times as much to fix a problem that could have been a quick fix.

One critical thing - make sure DSREPAIR shows that Time is in sync on all servers. Do this from the Master replica of [root]. That is the #1 priority for NDS. If you get that fixed, well, its a start at least.


Marvin Huffaker, MCNE
 
Hi Terry, yes there are only 4000 objects and the line is ISDN2 (128k). It is slow but the good news is that it synced ok overnight and the replica turned on. What was worrying was the info I was seeing in dstrace "the replica not on" warning about the root partition. This seems to be misleading but has also resolved. There is something weird with this network but we are fixing them one step ata time.

Marv, take a chill pill man! The NDS is not in a state of meltdown and Im not losing sleep over it. The problems were already here so they've been like this for a while. Time is synced ok and the replicas are all in sync. The culprit turned out to be a replica on another server in skulk that had been sat in that state or a while.

Thanks for your input guys.
 
good to hear

marv did make some very good points though - in terms of the server centric nature of these type of issue and indeed that quite often people do start to do multiple things that do produce a meltdown

looks like as you say one at a time and relax a while

dont 128k suck - hate the way it's 20 quid for a 4mb link into your house but come to business ands it's a mega bucks
 
Indeed. This company turns over 150 nillion a year and makes a clear progit of 55 million. Good news for the shareholders but the IT systems suck. At least they havent fallen of the Novell wagon, yet !
 
Sorry you feel that way Marv, but your reply was not very constructive. I can read and understand all the blurb on the Novell site explainig error codes and their fixes but when that fails I ask for advice. As it turned out it was not so serious and all was resolved by being patient - as you have so rightly advised in the past. Thankyou for your reply but I did not intend to waste anybodys time.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top