SQL Server Performance Forum – Threads Archive

What are Possible reasons for automatic failover?

I’ve a active active cluster enviornment with nodes OP1 and OP2 I suddenly see one node gets failed over to other node. I wanted to know what are the possible causes for the prod server failover to other node automatically? I did see in the even log but not able to see the proper / relevant reason for that. If there are space crunch in some server will it cause automatic failover? Thanks in advance guys…

There is a possible if there are space issues on one of the server, check the cluster & sql error logs for this reason. Satya SKJ
Moderator
http://www.SQL-Server-Performance.Com/forum
This posting is provided â€œAS ISâ€ with no rights for the sake of knowledge sharing.

satya any article / pointers / white paper for the space issues & reason for cluster failover?

You will find the cause of the failover in the event logs or the cluster and sql server logs.

http://support.microsoft.com/default.aspx?scid=kb;en-us;254321&sd=tech for info. Satya SKJ
Moderator
http://www.SQL-Server-Performance.Com/forum
This posting is provided â€œAS ISâ€ with no rights for the sake of knowledge sharing.

I’ve a active active cluster setup, very often i did observe one of the node gets hung. i come to know this hung issue when i try to TS in to the box, if the box doesn’t allow me to ts (no login screen appears) then i come to know the node is hung. What i want from u people – Is there any mechanism like monitoring i can setup which will tell me in advance the node is hung some thing like that ?

What does the event logs say when you get in after a reboot? "Hung" is a loose term. Something on the server is causing it to not respond, be it a faulty application that take 100% cpu or hardware error with memory etc. If the server is non responsive a standarad monitoring check with say a ping or a script that perform an action on the server will work.

I dont see any error mssg in the eventviewer related to memory / cpu / space. as i mentioned since i dint have an alerting system to inform server hang issue to my mail box, i m not able to see the exact time when it starts reaching in to hang stage. because of this i m not able to diagnoise the exact time and its related mssg in the event viewer just the approximate time a window of 3 to 4 hrs i do diagnoise. Could you let me know is there any separate event viewer for capturing cluster issues ? if so what is the option – (like my computer->manag -> eventviewer) whts the way to see cluster event viewer?

No, every information about server & cluster will be logged to event viewer. Satya SKJ
Contributing Editor & Forums Moderator
http://www.SQL-Server-Performance.Com
This posting is provided â€œAS ISâ€ with no rights for the sake of knowledge sharing.

]]>