cryptofreak.org cryptofreak home projects
contact about
Contact:


projects
News Agenda
Antera Antera
News Commentator
News fcreate
Linux Porting Linux Porting
mod-chal mod-chal
Quake III Quake III
News Zope
Contact: webmaster

From: Perry and Lorae Merritt (plmerritt, hypermall dot net)
Date: 2001.10.28 - 16.00 MST


X-Mailer: Microsoft Outlook Express 5.50.4522.1200

I put my comments throughout June's. I haven't read the doc yet.

Dan, there is stuff in here that you need to respond to.

Perry

----- Original Message -----
From: "June Mullins" <junemullins, earthlink dot net>
To: <antera, cryptofreak dot org>
Sent: Sunday, October 28, 2001 12:40 PM
Subject: Arch doc (again)


> Since we're meeting on Tuesday, I thought it was about time I got around
to
> making the updates we talked about last Tuesday.  Please check em out.
> I didn't do anything about the focus issue.  We talked about the fact
> the the
> document describes the general Antera architecture, but doesn't say much
> about the archiving appliance we are actually focussing on building.
> That's
> because the architecture is much more general, and is really independent
of
> any particular application.  However, the general focus means we don't say
> anything about particular tools, such as HSM support code; and we don't
> really say much about archiving at all.  Is this satisfactory????
>
> I assume we're meeting at 1:30 again.  Does anyone want to change the
time?
> Would it be more convenient if we met earlier or later?

1:30 is ok with me. Earlier, like 12:30 would be even better, but I'm
flexible.

>
> Remember other unresolved issues from last meeting:
>
> - Is there a problem with timeouts when we need to restore a file from
> tape?  Do we block on open,
>   and wait until the entire file is restored?  Do we block on read, and
> wait only until we can satisfy the
>   read?

Based on what I just read from Jay, I think we've got a problem with
blocking at all. I still think we need to test it (just becasue I hope Jay
is wrong). If clients are going to be timing out we've got to figure out a
way to keep that from happening. Maybe a combination of blocking the open
for 30 seconds or so and making the initial the initial reads slow until we
get the file loaded in. Maybe a slow open and blocked reads for the data
that is not migrated back to disk. Maybe...

> - Can we (should we) take advantage of HSM-aware applications?

And those would be??


> - Do we need dynamically expandable storage?  With RAID protection?

I thought we were going to be working ADML with MD/LVM, holding off on ARVM
until later.


> - Would it be reasonable to require the customer to have a tape library
> with more than one tape drive?

I think we're either going to sell them a complete solution, sell them a
disk solution if they have a library we can use, or sell software to a
VAD/VAR and let them build the solution.

>   Two tape drives would be useful for autorestore during disaster
recovery.


> - We need to know more about what the different management policies
> would be.

I think we can pick a couple to start with. (For sure, Dan and I can pick a
couple) After that I think we need to get customer requirements to buld what
they want/need.

> - What do we want to build for a prototype?
>

I think a proto is the ADML, with no persisted data, that allows files to be
backed up and HSM'd out and in. I think we have one policy, a full backup. I
think we do the backup, HSM a select (canned)group of files. Do an LS on
them to show that they look like they're still there. Open one of them to
show the library bringing the file back.

If we have time, based on the need for a proto, I think we add a virus scan
to the backup sequence.

>         June
>
>

--
This is the antera mailing list.  To unsubscribe, email
majordomo, cryptofreak dot org with message body `unsubscribe antera'.
Or, for more information, visit http://www.cryptofreak.org/.



This archive was generated by hypermail 2b30 : 2001.10.29 - 04.02 MST