[OpenAFS] Puzzler: lack of access to AFS files

[Christopher D. Clausen]

Rodney M. Dyer <rmdyer@uncc.edu> wrote:
> At 05:26 PM 12/12/2007, Jeffrey Altman wrote:
>> I disagree.  We need more resources for testing a broader range of
>> scenarios than we currently have available.  The performance
>> improvements must be implemented or you absolutely should go find
>> something else to use.
>>
>> If we can't get to the point where operations are as fast or faster
>> than NFS or CIFS and if we can't support all of the application
>> operations they support and if we can't scale to the number of
>> clients per server and requests per second that they can scale to,
>> you might as well go find something new.
>
> I understand this, however you need to realize where I'm coming from.
> We support professors who have research projects that run into the
> millions of dollars.

Perhaps some of these millions of dollars from these research projects 
can go into testing to provide a "better" AFS client that is both fast 
AND reliable.

> Many times these people don't know anything
> about where their data files are being saved when they choose
> "File->Save" from an application.  They expect it to work.  We need
> to be in a position to provide the "works" part.

Are you currently paying to cover any of the development costs for AFS? 
Do you have a support contract with any company specifically for AFS 
support?

> If they save a
> valuable data file from an application one day, then return the next
> and the application won't load it because of some random network
> change updated a few bytes here or there when the file was saved,
> what do we tell them?  "Oh btw, maybe you should keep a local copy on
> your USB keychain unless the AFS network fails?"  Most professors
> don't spend the extra time to run checksums on their files after the
> save.  This kind of thing doesn't cut it.  I'm the type of
> "professional" sysadmin who's willing to give up 10 percent of my
> speed for guaranteed delivery.  I'm not some young post high school
> geek who's got a job running a smallish home network and constantly

Some of us ARE young post high-school geeks who have jobs running 
smallish networks.  I thought a benefit of AFS is scalability?  What is 
wrong with scaling down?  Remember that some of us young post 
high-school geeks grow up to have jobs as "professional" sysadmins.

> boasts product x is faster than product y, and that's just uber cool
> because product y sux'ors!

I find that now is an appropriate time to post this link:
http://people.ccmr.cornell.edu/~mitch/afsvsnfs.html

But seriously, if AFS is at the point where non-professional geeks look 
at it and say "AFS rules!" then something has been done right.  Right 
now people just look at it and say "its not that bad" and then go on and 
look at other "cool" alternatives.

Many would-be AFS admins stop by and ask questions in the #openafs IRC 
channel.  Most of them go something like this:
* newuser1 has joined the channel
newuser1: Hi!  I heard that AFS can do replication.
afsadmin1: yes, but only for read-only data
afsadmin2: if you want real-time replication, you probably need to look 
at something else.
newuser1: oh? really? Too bad.
* newuser1 has left the channel

The other conversations involve those already using AFS and post 
high-school geeks who DO want to setup something cool.  The AFS 
community isn't going to grow if these people are insulted and 
discouraged from testing various new and cool technologies.

> I am happy with the speed improvements, and I hope we can continue to
> use AFS.  However I need to be able to look at people with a straight
> face when they ask about how well AFS works.
>
>      Speed?  Check
>      Scale?  Check
>      Functionality?  Check
>      Reliablity?  hrm...

I know this is isn't a useful data point, but to my knowledge, none of 
the AFS servers that I maintain have lost important data due to a fault 
in AFS.  Yes, some test data was lost, but that is exactly why a 
"professional" sysadmin runs tests in the first place.  Have you 
actually lost data?  Or are you just concerned about truthful warnings 
posted by the developers?  (Of course I realize that there is always the 
possibility that data is corrupted and one doesn't know yet.  Volunteer 
and help test new builds to help reduce these posibilities or fund 
development.)

I will also point out that a salesperson for a commmercial company isn't 
as likely to tell you that his/her company's product will not work in 
your situation.  The AFS community IS more likely to tell you the 
reality of the situation.  AFS is not better than filesystem y, at least 
not yet.

<<CDC