On 5/8/24 11:06, Lucio De Re wrote:
> There is much I would like to explain, but the problem I am attempting to 
> solve ought to have an obvious answer that I am clearly missing.
> 
> I can't seem to get a 9front workstation to mount a networked 9legacy fossil 
> service. The FS is a fairly pristine 9legacy installation, on a somewhat old 
> 386 platform. I did need to tweak various parameters on both side, but 
> eventually I got to the point where both hosts declare that the connection 
> has been established; now on the 9front workstation I get the message
>     "srv net!192.96.33.148!9fs: mount failed: fossil authCheck: auth protocol 
> not finished"
> I suspect the culprit is the lack of the newer "dp9ik" security on 9legacy, 
> in which case it would be helpful to know how to work around that.

Probably. Why not just temporarily disable auth checks for the fossil 9legacy 
machine?
Or perhaps just take a disk/mkfs backup and tar that. You really have chosen 
the most painful way of accomplishing this (which you seem to acknowledge).
Or just exportfs the root? There are so many ways of just getting the files.

> 
> Why am I mixing my platforms like this? Because the hardware on which I am 
> attempting to recover a rather large historical file system is split between 
> IDE and SATA and I have no hardware that can handle both disk modes and I 
> need to move information between the two media types. I am not describing all 
> the dead ends I tried, incidentally, that would take too long and really 
> expose my limited understanding.
> 
> It took almost a day to copy the Fossil cache (or lose a lot of the most 
> recent changes) and now I need (or at least want) to update the default boot 
> ("arenas") Venti configuration on a SATA drive which I can only access on 
> hardware I can't install 9legacy on. It's complicated and I'm sure there are 
> people here who would not find this so daunting, but that's where I am at. To 
> be precise, I need to change the Fossil default configuration (in the 
> "fossil" cache) so it points to the correct Venti
> arenas. I'll deal with the analogous Venti situation when I get past the 
> total absence of Fossil tools on 9front.
> 
> I guess I can port fossil/conf to 9front, but I'm not sure I have the stomach 
> to try that. Maybe now that I have raised the possibility...

It sound like you're trying to make this someone else's problem.
Being stuck in a hardware pickle when there are ample existing software 
solutions is not
a good reason to ask someone else to go out of their way to write software.

Fossil can be pulled in largely without modifications as I understand it,
I don't run fossil but some people in the 9front community do and it does
not appear to me that they've had issues with continuing to have it work
(other then fossil bugs itself).

> 
> I managed to share the Fossil cache through a NetBSD server providing u9fs 
> services, but that host does not have the capacity to store the Venti arenas, 
> nor can I really justify spending the amount of time it would take to pass it 
> between the 9legacy and 9front devices via NetBSD, no matter how I try to 
> arrange that. It does baffle me, though, that a NetBSD intermediary is more 
> competent than the two "native" platforms.

Are you blaming us for moving on from AES 53 bit keys that can be brute forced 
in an afternoon?
I have tried to open a dialogue for getting dp9ik on 9legacy a couple times 
now, when I had brought it
up I am told to write the patch. Something about being asked to spend the work 
to write a patch for 9legacy given
the historical context of why 9front exists does not sit right with me. So it 
wont be me, sorry.
Sure it sucks that things have drifted, but all our code is there, neatly 
organized out in to commits, if someone
wants to import our work it is readily available. However something tells me 
most people are just going to use 9front as is.


Good luck,
moody


------------------------------------------
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Tde2ca2adda383a3a-M6a4b734a4d1906cf86d9da88
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Reply via email to