On 02/11/10 05:53 PM, Frank Middleton wrote:
On 02/11/10 06:15 PM, Shawn Walker wrote:
On 02/ 9/10 03:36 PM, Frank Middleton wrote:
You can't continue a failed update as there's no reliable way to resume
a failed one. With that said, an update on a fast system (x86) should
take around 5-10 minutes since you've already downloaded all the files
for the update (and those should be cached).
Sorry if my terminology was incorrect. When I said resume, I
meant restart from some point (without having to download the
image update again).
Ah, then you don't have to worry about that.
If you don't delete the old BE you get a python stack dump (BE
already exists). If you delete it, it seems to delete the cache, so
If you get a python stack dump, that's a bug, and please file one at
defect.opensolaris.org if you can't find under under development -> pkg.
you are in effect starting all over again and the COMSTAR issue
wil cause it to fail again. Is there a way out of this catch-22 other
than going back to iscsitgt? IMO this is a defect in the packaging
system in that once the install process has begun and the BE has
been created, there's no way to restart it if there is some kind of
failure. Or is there?
I think there's some confusion here. The download cache is always
stored in the *current* BE before the new BE is created. While you will
see the client "go through" the download process -- it shouldn't
actually download files unless they're missing or corrupt.
Yes, we're aware that's slightly confusing, and hope to improve that in
future releases.
Cheers,
--
Shawn Walker
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
