Questioning my competence
Nov. 17th, 2006 06:16 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
trixtahYou know, I don't profess to know everything. *gasp!* And I have quite happily admitted all the way along that this whole Linux/Postfix thing has been a steepish learning curve for someone who has been a Windows admin for over 8 years. I've happily admitted it to my colleagues as well. However, while I'm fine with constructive criticism (most of the time, I'm not perfect, and it has been a hard lesson to learn over my life) and particularly suggestions about how to do things better, I intensely dislike being doubted in matters in which I feel fairly competent.
Today a slight issue came up with a kludge we currently have to put in place for users who have more than one mail alias. The current system does not play nicely with the Windows Active Directory, and only "sees" the default mail address for the user, and not any additional aliases. So if we need to cover variations like "steven.bloggs" for "stephen.bloggs", we have to put in a Custom Recipient and redirect the additional address to the correct one. It's annoying and shouldn't be necessary.
When I put Postfix in, it won't be necessary to do that kludge, since the address harvesting routine I'll have in place will gather all the aliases that might be associated with a mailbox. This will be much nicer. It's a script that runs at an hourly interval, picks it all up, does a bit of magic and then puts all the addresses on the Postfix server to be used for mail delivery lookups.
I sent an email to the accounts creation people today saying that when we switch over to the new system, we will have to find the existing kludges and tidy them up. Once it's done, everything will be nicely streamlined, and since there are only a couple of dozen of them, it won't take long. So, I get a return email from one of the VMS admins (who helps look after the existing mail systems, kind of, and whom I normally get on well with), totally questioning everything that I've put in place so far:
I admit that I could have been a little faster with the nuts-and-bolts doco that I need to give to him and the other Unix admin, but my design doco quite clearly laid out my aims (espcially removing the custom attribute), and sketched out the main methods I was going to use to achieve it (including the asynchronous address-gathering routine). Since the fucking email is my responsibility, and is what my position rests on, the nuts-and-bolts doco is only a courtesy anyway. If the mail is delivered, that's all he needs to know. Questions about the base design could have been asked months ago, not when I've finalised it (all bar the shouting).
The thing is, he's been really helpful up until now about the actual OS-based stuff (since he and the other guy are responsible for all the Linux servers), but now he starts throwing spanners in the works? I sent a fairly lengthy email responding to all his points. Politely. But I hope that he doesn't start feeling twitchy about the whole thing, because that could cause major problems. While he is a VMS/Linux guru, he knows nothing at all about Postfix, and I was hoping that was a very clear demarcation line. Well, we'll see on Monday, no doubt. I'm hoping it's just dorky geekboy cluelessness and not a "you can't play with my toys" kind of thing.
I can't wait for this fucking project to go live, actually.
Today a slight issue came up with a kludge we currently have to put in place for users who have more than one mail alias. The current system does not play nicely with the Windows Active Directory, and only "sees" the default mail address for the user, and not any additional aliases. So if we need to cover variations like "steven.bloggs" for "stephen.bloggs", we have to put in a Custom Recipient and redirect the additional address to the correct one. It's annoying and shouldn't be necessary.
When I put Postfix in, it won't be necessary to do that kludge, since the address harvesting routine I'll have in place will gather all the aliases that might be associated with a mailbox. This will be much nicer. It's a script that runs at an hourly interval, picks it all up, does a bit of magic and then puts all the addresses on the Postfix server to be used for mail delivery lookups.
I sent an email to the accounts creation people today saying that when we switch over to the new system, we will have to find the existing kludges and tidy them up. Once it's done, everything will be nicely streamlined, and since there are only a couple of dozen of them, it won't take long. So, I get a return email from one of the VMS admins (who helps look after the existing mail systems, kind of, and whom I normally get on well with), totally questioning everything that I've put in place so far:
- The address harvesting shouldn't be asynchronous, we should be doing dynamic lookups. [Leaving aside the fact that I wouldn't be able to do my re-addressing magic that way, the load on the Postfix server and the AD and the delay while a remote query takes place for each message would be ridiculous.]
- Did I know that AD only lets you return 1000 records at a time during an LDAP query? And yunno, we've got 4000 users and all, so I might be missing email addresses with my address harvesting. [Uh, yes, since I haven't been a VMS admin for the last 8 years, I've been a Windows admin, and I do know how Windows/AD works, thank you. And my Perl script can do as many iterative queries of 990 records as necessary, while it dumps the results in an array for further processing - I certainly have all the email addresses.]
- We would have to get rid of a custom attribute in the AD that currently facilitates mail delivery. Oh, and it's linked to other stuff which has nothing to do with mail. [My process would ignore that attribute. Since it was specifically created to facilitate mail delivery, it shouldn't have been linked to anything else. Since it requires an extra step to create the attribute with the current account creation process, wouldn't it be better if we can now ditch it?]
I admit that I could have been a little faster with the nuts-and-bolts doco that I need to give to him and the other Unix admin, but my design doco quite clearly laid out my aims (espcially removing the custom attribute), and sketched out the main methods I was going to use to achieve it (including the asynchronous address-gathering routine). Since the fucking email is my responsibility, and is what my position rests on, the nuts-and-bolts doco is only a courtesy anyway. If the mail is delivered, that's all he needs to know. Questions about the base design could have been asked months ago, not when I've finalised it (all bar the shouting).
The thing is, he's been really helpful up until now about the actual OS-based stuff (since he and the other guy are responsible for all the Linux servers), but now he starts throwing spanners in the works? I sent a fairly lengthy email responding to all his points. Politely. But I hope that he doesn't start feeling twitchy about the whole thing, because that could cause major problems. While he is a VMS/Linux guru, he knows nothing at all about Postfix, and I was hoping that was a very clear demarcation line. Well, we'll see on Monday, no doubt. I'm hoping it's just dorky geekboy cluelessness and not a "you can't play with my toys" kind of thing.
I can't wait for this fucking project to go live, actually.
