Comments for [digital:meditation]

Comment on cyrus imapd: shared folder hierarchy recovery by kr

kr — Tue, 02 Feb 2010 20:20:05 +0000

Folks;

sorry to say I can’t offer you with any more in-depth insights on that… actually, given this approach (copying things back using mc) was likely to fail, I would have been in for a more, well, “brute-force” approach to things – copying the “restored” folder to a quiet place and hack up a small application utilizing JavaMail and the cyrus administrative login to traverse the restored folder hierarchy, looking for mailtext and appending it to IMAP through Java. However, a few thoughts to add, nevertheless:

- Did you look at recovering files / folders using the right (file system) permissions? This is something I stumbled across in my very early cyrus experiments (version 1.5 IIRC).

- Tried re-creating any of the mailboxes in question manually (i.e. using cyradm), copying mailtext and stuff off the recovered structure there and doing a cyrreconstruct afterwards?

- Do you have an idea how loss of mails / structure actually did happen in your system? In my environment, “fortunately” it just was a single user who accidentially moved the mess to Trash, so despite being “wrong” from an operation point of view, it still was “the right thing to happen” from the server point of view, not a malfunction or software failure of some sort…

Comment on cyrus imapd: shared folder hierarchy recovery by Jason

Jason — Tue, 02 Feb 2010 15:52:42 +0000

I am in the same boat as you bsh. I wish reconstruct would just do what they say it does. If anyone has any incite to this issue it would be greatly appreciated.

Comment on Sun, Oracle, visions achieved and points missed? by kr

kr — Mon, 25 Jan 2010 20:03:09 +0000

Well yeah, I don’t really think the philosophy in itself is all too bad, but then again, at first: The way they have been living up to this philosophy just hasn’t been really concise even the last couple of years… Just talking about “OpenSolaris” (the kernel) vs. “OpenSolaris” (Suns very own distribution based upon the OpenSolaris kernel). In this situation, why not just simply open-source everything of it and, all along with the “you can do it, so do it” – philosophy, let someone pick things up, let a couple of OpenSolaris distributions grow (and maybe even surpass GNU/Linux ones sooner or later – which should be easy given the OpenSolaris kernel comes with quite a bunch of good technology). By now however, there is “OpenSolaris built on top of OpenSolaris”, and “open OpenSolaris” distributions (Nexenta, StormOS, Belenix) seem to struggle both lacking developers and users. And, on the other side, I mean, after all: As we see, adhering to a good philosophy is fine but in the end a company needs to earn some money each month not to run out of business… looking at the last two years at the very least, it seems just this is where Sun, overally, failed, and maybe things would be different now if, asides a good philosophy, Sun had taken a chance to actually make money out of building just one of these “batteries included” solutions. Oh well, we will see what future holds, no doubt about this… :)

Comment on Sun, Oracle, visions achieved and points missed? by Markus Schmidt

Markus Schmidt — Mon, 25 Jan 2010 19:49:35 +0000

Well, I agree to your ‘missed options’, but as I see, the main reasons for that is SUN’s philosophy. Maybe it’s because they are a ’students company’ as you name it, they don’t see it as their job to glue all the technologies together. We, as a Sun Partner, asked severals times for integrated Solutions, the answer has been clear: “You’re a Sum Partner – build it” Last that happend in the virtual Desktop Field: No Windows (or Linux) Company in the World is going to implement a Sun Ray Architecture and maintain Solaris (ok, now it runs also on Linux) know how just to get the backend for some Thin Clients up and running. So basically it whould be no Problem for Sun to come up with a ’sun Ray Appliance’ – Solaris, SunRAy Software, Webserver, little bit of Logic & Interface, and ship it. We got the answer: “You can do it, so do it”.

Well, things seem to change, lastly seen with the OpenStorage stuff, maybe in the future with Oracle, but I have some doubts (as far as rumors go, there will be hard times for the Sparc Architecture, disegarding the fact that this architecture is quite superior to he x86 for a oracle database, at least in our customer base, Oracle seems to don’t like that.

Well, we’ll see….

Comment on cyrus imapd: shared folder hierarchy recovery by bsh

bsh — Fri, 22 Jan 2010 15:07:27 +0000

Hi
I’m in a similar boat, where all our shared stuff disappeared after years of reliable operation of cyrus. I have weekly backups of the mail storage folders, so I don’t even have to find out what goes where, just delete the maildir and extract from the tarball, and…. it doesn’t work. :( Reconstruct doesn’t help it either. All the mail files and folders are there, yet it doesn’t even see them. bah! :(
Still fighting with it.

Comment on eclipse+maven2: still a rough ride… by kr

kr — Thu, 17 Dec 2009 11:01:09 +0000

Yes. I tried. And cursed. And gave up on that idea altogether. Building Eclipse RxP applications using maven simply seems impossible (for that matters, building Eclipse RAP applications for the sake of deploying them as .war files generally is a tedious process no matter whether or not maven2 is involved as soon as you want it to happen headless… :/ ).

Comment on eclipse+maven2: still a rough ride… by markus

markus — Thu, 17 Dec 2009 10:51:58 +0000

And the real pain is just about to start => did you try to maven-build the RAP-part of your application too?

Comment on white-out? by kr

kr — Mon, 19 Oct 2009 13:09:47 +0000

Merci. :) Still gotta do some work, though…

Comment on white-out? by Ralf

Ralf — Mon, 19 Oct 2009 12:45:49 +0000

Looks nice though. ;)

Comment on a strange kind of cross-over: Botany Bay – “grounded” by kr

kr — Tue, 06 Oct 2009 05:53:13 +0000

@Stephan: You’re welcome, really enjoyed listening to the album. ;)