Swapinfo High (100%)

[Akif_1]

Hi Bill,
I go through with hp.docs ...... http://docs.hp.com/en/B2355-60105/dbc_min_pct.5.html?btnNext=next%A0%BB which shows me different picture of dbc_max_pct not to disturb or do calculation before making mind to change dbc_max_pct or dbc_min_pct
and other result of UNIX95 enclosed ......
I appreciate ur help and need more to close this thread...


Thanks,

[Michael Steele_2]

From your vmstat report which indicated zero idle time and your unix95 report, ora_d000 and ora_s000 are you're two biggest cpu consumers.

UNIX95=1 ps -ef -o pid,ppid,vsz,sz,cpu,args|sort -k3nr|more

6124 1 102464 25616 193 ora_d000_NPRD1
10770 1 102976 25440 72 ora_s000_NPRD1

However you've also got 14 NFS daemons up and running and that's a lot of NFS activity. Have you investigated NFS reports?

As I understand it DB1 is the only node out of four that are reacting poorly in this serviceguard environment. How is your serviceguard cluster mapped into these nodes? Do you have 14 NFS daemons up on these other nodes? Is the speed and duplex correctly set for DB1 and different for the other nodes?

Is oracle relying on nfs mounted filesystems?

These are the only configuration issues I can see in the O/S. Everything else points to the Oracle side of the server. But in reviewing the UNIX95 report I don't see a lot of memory problems. So I'd concentrate on what's different with DB1 and the other nodes and start comparing configurations.

What's your lan speed and duplexs on every node?

[Akif_1]

Dear Michel,

How to check this duplex speed of nic as i go through with my cfg2html file displays me this output in all the four servers. (Technology Specific:
Speed/Duplex = NO LINK AUTONEG )

Cheez:)

[nanan]

Hi Akif

Use lanadmin
lanadmin -x

You can get the PPA NO you concerned with lanscan

Regards
nanan

[Akif_1]

Thanks for your valuable support