Migrating Exchange

I’m in the process of migrating an Exchange server to new hardware. At the same time the Exchange organization name and structure were changed and reorganized, so the only way to move the existing mails between the servers was to use exmerge.

I was scratching my head (and pulling my hair) for several hours with strange errors. Extraction of many mailboxes failed with errors such as “MAPI_W_PARTIAL_COMPLETION” in the log files. I killed anti-virus and anti-spam services on the server but that made no difference. So I increased logging level and got a new error message: “Error opening message store (MSEMS).” With the “help” from that error message (and Google) I found a MS support page that resolved the problem; the server’s Administrative account had “deny” permissions set for Receive As and Send As.

So far so good, now all the smaller mailboxes seem to be extracting fine. But there are five mailboxes whose sizes range from 2.7Gb to over 5Gb. The PST files have a 2Gb limit, so now I’ll have to do multiple runs for those mailboxes with date ranges, such as pre-2005, 2005, 2006, and 2007 (I found some tips on date-limiting extraction here)… that should break the large mailboxes to small enough pieces so that the corresponding PST files will be below the 2Gb limit. I find it strange that this is the only method of transferring data from an Exchange installation to another (when not in the same forest). If the data store file could not be migrated, why not offer some kind of binary transfer file format that could be of any size? Or—at least—why not offer an automated splitting of the output PST files; when one reaches 1.9Gb, it would automatically be closed and a new one opened with a sequential number after the name. The current way is awfully cumbersome, especially considering that it’s rather common for users to have large mailboxes in Exchange that doesn’t have the size-limitation of the PST files.

And what’s with the X.500 sender addresses after the move?! Considering that Microsoft periodically releases new versions of Exchange, and that they assumably want people to upgrade, why to make upgrading so painful? All this would be so simple in the UNIX world, for example, with Postfix + Dovecot.

Do you send feedback to software manufacturers? You should!

I rather frequently send feedback to the manufacturers of various software packages I use. Sometimes my suggestions or comments are received enthusiastically, sometimes they’re merely acknowledged, and yet sometimes I never hear back from the software manufacturer in question. In most cases, however, the maker of the software is happy to receive feedback on their creation, and also in many cases if there was a problem or a feature request, they take action on it: the described problem is corrected, or the requested feature is added in the next release (or sometimes sooner with a patch).

How the comments are received seems to often correlate with at least two factors: whether the feedback was solicited, and the size of the outfit. Perhaps a bit paradoxically smaller the organization, often more rapid the response. This has probably its roots in a more developed “ownership” of the product in question by a specific person or a small group than in large companies in which various developers might develop a segment of a software, but none of them feel like the software would be specifically “their” product. Larger organizations often say they appreciate and welcome feedback but—perhaps due to the large number of users sending them feedback—one rarely hears back from them (beyond an automated “thanks”). Looking at the subsequent version of their software it’s often equally difficult to tell whether my comments specifically resulted in anything. At least the larger software companies like Adobe tally all the feedback they receive to identify the most requested features, and then provide something that covers as many requests as well as possible (in case of Adobe I liked their extensive push to collect opinions from the Photoshop users while CS2 was under development; they may have repeated the process with CS3). While giving a company with large number of customers a better idea how to prioritize their product development, this process obscures the direct relationship between a specific feedback message and a future feature addition or a bug fix. However, as long as the commented problem is somehow circumvented or the requested functionality achieved—even though it might not be implemented exactly as was suggested in the feedback—I suppose it’s not so important to know if it was really my message that made the difference (either directly, or as a “vote” for a new feature or fix).

Nevertheless I think it’s a good practice to provide feedback as it generally results in improved software. Often the developers may be “too close” to their product, and as such may have missed something that appears very obvious to the user.

HP responds

After I wrote about HP’s lame service policy a bit over a week ago, I forwarded a link to the article to HP’s CEO using a form on HP’s site. A PR rep from HP called me on July 4 (!) to provide more info about their service policy. It didn’t appear she had read my blog post, so I gave her a quick run-down of why I was unhappy with their service policy—that a $299 fee to replace a tiny plastic part on my laptop’s keyboard seemed outrageous, and that the “added value” items included in the service carried no added value to me.

Turns out HP does sell spare parts. Not the plastic supports individually, of course, but the entire keyboard module can be purchased for fourty-some dollars (about $58 with tax & shipping). That’s starting to be in the ballpark of how much I might expect to pay to fix a broken keyboard (it would be great if the actual part that was broken could be purchased for a few bucks, but that’s too much to ask or expect from a large corporation). The reason or rationale for why the service rep didn’t offer such option to begin with is, according to the HP’s PR rep who called me on Wednesday, that relatively few people are comfortable to carry out such repair and/or capable to do so without further damaging their laptop… which then often gets sent in to HP to be put back together—a procedure much more expensive [for HP] than replacement of just the keyboard. Naturally a self-attempted service on a laptop that is still under warranty also voids the warranty.

Fair enough. I can see where they’re coming from. I would still like to see their service reps to mention the availability of spare parts, at least when asked. I did ask, and the service rep told me that spare parts were not offered—but perhaps she didn’t know (or was not allowed to say) since the corporate goal at HP appears to be to discourage the bungling masses from attempting to service their laptops themselves, no matter how minor the problem, and even when (and perhaps especially when) a laptop is already out of the warranty (so that a user really has nothing to lose.. except for $299).

For your future reference, here’s the phone number to HP’s spare parts service: (800) 227-8164.