dsrepair woes

[terry712]

all servers netware 6 sp3 edir 8.7 - vanilla so i know it's a minor version out

when i try to run a full unattended repair on any of the 5 that hold a copy of root - it bombs out with an error (the old insufficient space one)

even if you try to run with -av - still does nothing
you can run a normal repair from the repair db screen
and it can be run on any server that holds any other replica except root

any thoughts - i'm a bit confused as my way of thinking is during a repair like this - its really a server centric repair which would suggest corruption on that partition

havent try dsrepair when unconnected to network or with a -ndb and then try

any thoughts or seen it
i dont want to upgrade to 8.7 whatever it is until all ok
tcpip patch still needs to go on as well

 

[marvhuffaker]

Terry, did you check how much space you had available on SYS:? I can't really tell from your post.

I only ask because the space requirements vary depending on the type of repair you run.

Also, what are the states of the partitions, did you just recently add/change a parition or replica? What do you see when you do a SET DSTRACE=ON, SET DSTRACE=+S, SET DSTRACE=*H... Do you see errors or is it clean?



Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[terry712]

it's affecting 5 server - all holding master replica's
space on all approx 2 -3 gb spare.
no replication changes for a few months. no stuck obits on any of them.

on the dstrace's - i'm getting the expected 698 skulkers but i'm also getting some 625's - always pointing back to one server but from different reps.

the server is pingable , resolevable , shows in slp services and display servers and if you do a timesync check.
also if you do partition status - the ds is staying in sync but you get 625's against some

the server involved is a dl380 and i did recently upgrade the firmware on everything on it and also applied softpaq 6.41

i'm going to backrev the softpaq to 6.3 incase it's a funny nic issue and it's dropping packets or something - even if it is this though cant understand why cant dsrepair on others but as you say the 625 will need to be resolved first then worry about next

 

[marvhuffaker]

625 errors aren't normal. You shouldn't be getting them. That means that the servers aren't synching. I sense that you have some deeper nds issues that need to be resolved.

When you say they all hold master replicas, can you expand on that?

You should troubleshoot the 625 errors first, there are plenty of Novell TIDs on that topic.




Marvin Huffaker MCNE, CNE
Marvin Huffaker Consulting

http://www.redjuju.com

 

[TheBoySmith]

There is an issue with softpaq 6.4 and n100 cards. Not sure if it is related to your hardware.

The fix is to backrev the n100 driver. Loads of comms issues with the 6.4x one.


Lee Smith
Associated Network Services
Snr Systems Engineer
Lee.Smith@ANSPLC.Com

 

[terry712]

bacrevved to 6.3 and ds is clean now

it is the n100 but backrevved complete softpaq rather than have a mixed bag


still can unattend dsrepair

what i mean is the 5 servers that i cant run it on - each hold a copy of the root partition

 

[TheBoySmith]

You only needed to do the n100 but no great shakes doing the full backrev.

To have a need to run a full unattended you moust have big probs. I suggest you try running each option individually.

My initial thought is - it is having an issue with repairing network addresses.

We need to find out which of the unattended options is causing the problem.

Can you run a local repair on each of the 5 servers?

Can you run a repair network addresses on each of the 5 servers?



Lee Smith
Associated Network Services
Snr Systems Engineer
Lee.Smith@ANSPLC.Com

 

[terry712]

the addresses was my first thought and i did that - but to be truthful i think they were ok

a local repair runs no probs - so i'm not overly concerned
just curious - but ds etc all seems ok so probably worry about it later

i have seen it before when you couldnt run a repiar and it's to do with faulty checking of space - you do a dsrepair -av and runs fine

but not in this case