Windows 2003 Server DFS and SANs
I have a couple of questions here. I have been in my current position for about 5 months and there was and still is much to clean up here. First and foremost, I have many single points of failure that I need to make redundant. I am hoping that DFS can help me with this.
Here is my first challenge. Both of my AD servers are running on W2K3 server - not R2! I don't have any experience with non-R2 W2K3 server and DFS/FRS seems to work differently or at least has a different GUI. I have increased storage on both servers so that I can setup replication for our 'public', 'users' and 'dept' shares. However, when I start FRS process, it quickly fails (event 13508, 13544, 13552) after replicating only a small percentage of the files, and then never recovers. I have goggled this until there is nothing new to read on the errors I am getting. So, here is my question. Will upgrading these two servers to R2 resolve this issue, or is there something else screwing me up here?
Second question: If I have two servers that both have access to two SAN drives, with both of them sharing those drives with the same share names, can I configure a DFS share that has them both a root shares, but not enable FRS (no need to) - that way, if one server is down, the other will automatically services any requests to the DFS share.
Thanks,
David
Here is my first challenge. Both of my AD servers are running on W2K3 server - not R2! I don't have any experience with non-R2 W2K3 server and DFS/FRS seems to work differently or at least has a different GUI. I have increased storage on both servers so that I can setup replication for our 'public', 'users' and 'dept' shares. However, when I start FRS process, it quickly fails (event 13508, 13544, 13552) after replicating only a small percentage of the files, and then never recovers. I have goggled this until there is nothing new to read on the errors I am getting. So, here is my question. Will upgrading these two servers to R2 resolve this issue, or is there something else screwing me up here?
Second question: If I have two servers that both have access to two SAN drives, with both of them sharing those drives with the same share names, can I configure a DFS share that has them both a root shares, but not enable FRS (no need to) - that way, if one server is down, the other will automatically services any requests to the DFS share.
Thanks,
David
Lofty Worm
In relation to your second question, the answer I think is 'no'. The reason is that only one server may access any single disk at one time. If you wan to present a disk to 2 servers, you need to set up a cluster. Of coarse you could set up the disk to be available to both (wiring), but then when there is a server failure, you would have to manually intervene to actually connect the disk so the data could be shared, AND there is the worry that windows may try to 'help' and load the disk anyway, causing disk and data failures.
As for the first, there are no guarantees in life, only death and taxes. My experience has shown that upgrading in hopes of fixing an issue has failed 9 out of 10 times for me. From your description, I would be looking at the network connects. If FRS starts and is working and fails half way through, I am thinking the network lag or latency may be causing to many errors, and windows is saying 'enough'.
There are some specifc FRS logs, but I woul have to dig around to find out were they are hiding.
There are some specifc FRS logs, but I woul have to dig around to find out were they are hiding.
ASKER
The IBM Storage manager software allows for configuring a group of hosts to access the same logical drives. There is nothing in documentation on how the system deals with any possible contention with resources on the logical drives. I assume, since the OS accesses the resources through a driver that talks to the IBM Storage System that the driver would handle this issue.
As for network issues, they are both connected to the same gigabit switch. I would assume that isn't the issue.
David
As for network issues, they are both connected to the same gigabit switch. I would assume that isn't the issue.
David
It has been my experience that drivers do not take care of the disk issue. But, is just my opinion, and it would not be the first time I was wrong.
Have you checked the speed and duplex on the NIC's? Are the switch ports reporting any errors. Perfmon will also tell you if there are errors (and monitor lots more), but run it during a test of FRS.
If you want to look at the FRS logs, try this
http://technet.microsoft.com/en-us/library/cc978208.aspx
Have you checked the speed and duplex on the NIC's? Are the switch ports reporting any errors. Perfmon will also tell you if there are errors (and monitor lots more), but run it during a test of FRS.
If you want to look at the FRS logs, try this
http://technet.microsoft.com/en-us/library/cc978208.aspx
ASKER
I just found the info I was looking for concerning hosts sharing LUNs in the SAN via a Host Group. This function is only for clustered servers - which would eliminate the need for DFS.
I am currently rerunning the FRS initialization and gathering logs and performance data.
I am currently rerunning the FRS initialization and gathering logs and performance data.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.