UPDATE: We have now released the hotfix that corrects this issue for Exchange 2003 SP2. The article is: 930241 The Exchange 2003 database does not mount, and event IDs 9518 and 9519 are logge...
Updated Jul 01, 2019
Version 2.0
Blog Post
So those of us running Exchange 2003 on Windows 2003 and running A/P clusters...
I'm reading from various posts that you should patch the idle node first for both Windows and Exchange, reboot it once all that's done, then slide services over to that guy, then patch up what was the active guy...
Based on what I'm reading though, I'm wondering if I should be patching my live guy first, so that I can fall back on my un-666'ed server if my stores don't come back. Does this make sense?
Of course, that only works if this patch (the OS and 666 patches) aren't doing anything to the store that an unpatched node can't understand and mount. Seems to make sense to me... that's why we need the tool to go in and invasively fix the stores data, because the patch itself isn't doing anything to the store or store information, just the binaries of the "program" of Exchange and Windows. So if I turn off my passive node, patch my active server, and it all goes to hell - can I just shut down my active server and power on my passive and make him the new active?
I realize that won't fix the issue, but at least I'm not left with an unmountable store for an extended amount of time...
I'm reading from various posts that you should patch the idle node first for both Windows and Exchange, reboot it once all that's done, then slide services over to that guy, then patch up what was the active guy...
Based on what I'm reading though, I'm wondering if I should be patching my live guy first, so that I can fall back on my un-666'ed server if my stores don't come back. Does this make sense?
Of course, that only works if this patch (the OS and 666 patches) aren't doing anything to the store that an unpatched node can't understand and mount. Seems to make sense to me... that's why we need the tool to go in and invasively fix the stores data, because the patch itself isn't doing anything to the store or store information, just the binaries of the "program" of Exchange and Windows. So if I turn off my passive node, patch my active server, and it all goes to hell - can I just shut down my active server and power on my passive and make him the new active?
I realize that won't fix the issue, but at least I'm not left with an unmountable store for an extended amount of time...