Here's an error message we got from Amanda during a recent restore attempt:

warning: restore program for /slocal/amanda/bin/ufsdump not available.
amrecover couldn't exec: No such file or directory
problem executing restore
amrecover: amrecover: pipe data reader has quit: Broken pipe

First the background: the filesystem that we were attempting to restorea file from had recently been migrated from our old fileservers to oneone of our new Solaris fileservers ;the backup we were restoring from had been made on

One of the interesting things that the Unix Heritage Society does is that it lets us go backand see just what a remarkable thing Seventh Edition Unix was, how much it was an aesthetic flowering and reallycreated a lot of what we now think of as 'Unix'.

It does this not so much by letting us going back to see V7 as it was(by modern standards, V7 is nothing terribly impressive and is in someways rather primitive). Rather, it does this by letting us go

It recently struck me that one of the things that the universitywebmail phish attacks demonstrate is thatreputation based antispam systems are now dead. The university webmailattacks aren't just a few previously good sources going bad, which hashappened before; they're a systemic, broad attack on a whole class ofsystems that previously had a good to great 'spam reputation'.

(Well, this exaggerates the situation somewhat. There are two aspects ofreputation based antispam systems; you can attempt to blacklist placesthat are

In theory, iSCSI has a feature called 'multiple connections persession', commonly abbreviated as MC/S. In iSCSI terminology, a'session' is a single logical connection between an iSCSI initiator andan iSCSI target, and a 'connection' is a TCP connection. MC/S lets asession be composed of multiple TCP connections, each of which can use adifferent set of IP addresses and thus a different network path.

(In practice, apparently very few iSCSI initiators and targets actuallysupport MC/

One of my blogging flaws is that when I write up an entry, I can be soclose to the issue that I leave important things out because they are'obvious' to me. I did that with my recent entry about Firefox 3's handling of self-signed SSL certificates: I skippeddescribing exactly what makes Firefox 3's approach so wrong.

Firefox 3 is not wrong-headed for being cautious about self-signed SSLcertificates, because there are real concerns with them in practice. Firefox 3

I was a little harsh on Firefox 3's handling of self-signed SSLcertificates in the last entry , and in fairness Ineed to say that this is a genuinely hard problem for browsers, at leastfrom a security perspective. Let us talk about just how this is so.

First, let us start with my previous basic model of how https could work. Browsers always use SSL if possible, and theyprotect against straightforward man in the middle attacks by rememberingthe SSL certificate they saw last time, or

There has been a certain amount of uproar about how Firefox 3handles self-signed SSL certificates ,and a certain amount of attempts to justify it . I disagreeviolently with the attempts to excuse the behavior, because whatit comes down to for me is that Firefox 3 has made it more attractive to have no SSL certificate at all than to have a self-signedone. This is both insane and inane; it does nothing to furthersecurity on the Internet, and it has basically nothing to do withthe sorts

One of the periodically proposed spam solutions is that someone willissue certificates to people and if they misbehave, the certificateswill be revoked. One of the many problems with this idea is what I willcall the corporate identity problem .

The problem with corporate identities is twofold, or perhapsthreefold. First, corporate identities are much like Internetidentities : they provide only positiveidentity, that person A is associated with corporation B, not negativeidentities, that corporation B is not associated with person A. In factthere are a

To summarize my earlier entry , the problem withinitrds is that they are both fragile and opaque. To improve them,one needs to tackle both issues.

The easy way to tackle the fragility is to make initrds both morecomprehensive and more dynamic. Since one way initrds break is by notincluding necessary kernel modules, simply include nearly everythingby default (you can exclude modules that are clearly not needed forbooting, like sound drivers); the extra disk space and boot-time memoryspace needed is likely to be