Re: Latest recommendation for asmlib + number of LUNs
From: Justin Mungal <justin_at_n0de.ws>
Date: Thu, 15 Aug 2013 05:30:56 -0500
Message-ID: <CAO9=aUxmyRa+2wzz9t0x1_p+mB-_ZEZ=UqaofsX=OotyYgffRg_at_mail.gmail.com>
I'm with Niall on this one. All of our RHEL 5 deployments use ASMlib; constantly updating the ASMlib kernel driver (and making sure one actually exists for whatever kernel you're planning on upgrading to) is a pain. This may have changed... but when RHEL 6 came out, Oracle, in their typical infinite wisdom, thought they could hegemonistically convert us all to OL by dropping ASMlib support for RHEL 6. Instead, we went with udev for persistent device naming/permissions and haven't looked back since. OS-level maintenance is much simpler and at most all I have to do now is a GI+RDBMS relink.
Date: Thu, 15 Aug 2013 05:30:56 -0500
Message-ID: <CAO9=aUxmyRa+2wzz9t0x1_p+mB-_ZEZ=UqaofsX=OotyYgffRg_at_mail.gmail.com>
I'm with Niall on this one. All of our RHEL 5 deployments use ASMlib; constantly updating the ASMlib kernel driver (and making sure one actually exists for whatever kernel you're planning on upgrading to) is a pain. This may have changed... but when RHEL 6 came out, Oracle, in their typical infinite wisdom, thought they could hegemonistically convert us all to OL by dropping ASMlib support for RHEL 6. Instead, we went with udev for persistent device naming/permissions and haven't looked back since. OS-level maintenance is much simpler and at most all I have to do now is a GI+RDBMS relink.
However, I will admit we'd be using OL if it were my choice. We probably will be one day in the future.
-- http://www.freelists.org/webpage/oracle-lReceived on Thu Aug 15 2013 - 12:30:56 CEST