@Phil ya

@JackDouglas hmm

@Phil yup

@JackDouglas hmm hard to tell with asm

@JackDouglas no, they're not

well, now I got myself confused there

looks like roland just copied the blog word for word - dba.stackexchange.com/a/42471/10064

@rfusca Linux?

It's nearly always IO

@Phil ya linux

thats what I thought

my boss wants to try restarting the db, so we're going to do that

I want to try moving the redo logs to another volume

@Phil i'm just not sure how at all to do a good IO test between the asm and raw dev of it

Have there been any software changes recently that have increased commit rates? and could such a process be just logged in to the one node?

@Phil nope, definitely nothing new

@rfusca You can't really, as it's writes

Are the storage muppets seeing anything?

this did lead me to find several bad offenders of 'Commit on every loop iteration' and move the commit outside the loop. It helped, but not enough. But that's been there for a long time

How is the storage presented?

@Phil lol if you're referring to the the guy in charge of the SAN - he was 'let go' a few months back and not replaced yet. His coworker is useless with SAN stuff

@Phil its a EMC Clariion and the LUNs are available in asm...does that answer your q?

Sorry, I meant what is the transport? Is it over ethernet? Could be packet loss

@Phil fiber

Unhappy about that

@Phil :(

well i opened an SR with Oracle - but not expecting much

Is it a shared disk array or dedicated?

@Phil shared with 1 other low volume app

@rfusca Right. The problem you have is that they will ask you to look at the storage. It's possible, and usually the case, that the spindles that have been happily servicing your sequential log IO have been joined by a load of random IO from your other apps

Do you have access to the EMC tools?

@Phil I think so but I don't really know anything about them :/