0

I have a Windows 2008 R2 failover cluster server. I am going to move one of our window services onto this new server. The service writes some trace information to a log file on the local harddrive. This will become a problem when it is moved to cluster server when cluster A become unavailable and cluster B takes over and now there are 2 places where I need to look for log files. Is there a way to make sure regardless of which cluster is on, I get one complete log file?

I have been researching this and there is something called DFS replication but i was wondering if there is something better that works with failover cluster...

I prefer not having to update my code. I can specify it to write log files to a different location by changing app.config file but no code change...

user156144
  • 115
  • 3
  • Can the log not be written to shared storage? – Shane Madden Oct 31 '13 at 00:12
  • That is a good point. Assume there are only 2 machines total. Putting it on a traditional windows file share on one of the 2 machines would defeat purpose of using failover cluster – user156144 Oct 31 '13 at 00:14
  • Assuming that you're implementing the service in an active/passive fashion and that it will be running on only one node in the cluster at a time then I don't see how this is an issue. If node A is the active node then you know the current log file is on node A. If node B is the active node then the current log file is on node B. I don't see why you need to "look" for the log file. You'll always look for it on the active node. Is there some logic that is escaping me? – joeqwerty Oct 31 '13 at 00:44
  • you are right that log files will be written to the active node. What I want is to have one log file that shows contents of both log files so each time i need to anaylze them, i don't have to merge those 2. Merging is more difficult than you think if you want chronological order – user156144 Oct 31 '13 at 01:16

1 Answers1

3

Normally a failover cluster will have some type of external shared storage like SAN or DAS. If your service is a clustered service, you should associate a shared volume with it and write your data to that. This way there is only a single copy of the data, and if your service fails over to the other node, the volume goes with it.

This is generally how clustered services work.

phoebus
  • 8,380
  • 1
  • 31
  • 30
  • thank you for the answer. Our set up is to have 2 nodes in different buildings in case one of the lose power. Our SAN is per building specific so we don't have external shared storage. Is there a way to do this with 2 servers and no external storage? – user156144 Oct 31 '13 at 18:40
  • @user156144 If you have another server you could expose some storage over iSCSI to the two servers, or you could have the service write to an external location like a file share. Note that in order to properly run a two-node cluster you will need some kind of external quorum witness such as a shared disk or file share, so you could use whatever storage that sits on. – phoebus Oct 31 '13 at 20:25