> From: Yaron Zabary <yaron AT ARISTO.TAU.AC DOT IL>
> Reply-To: EMC NetWorker discussion <NETWORKER AT LISTSERV.TEMPLE DOT EDU>, 
> Yaron
> Zabary <yaron AT ARISTO.TAU.AC DOT IL>
> Date: Sun, 1 Feb 2009 18:02:23 +0200
> To: <NETWORKER AT LISTSERV.TEMPLE DOT EDU>
> Subject: Re: [Networker] 7.4.2 -> 7.4.4 or 7.5 - upgrade
> thoughts/recommendations
> 
>    My experience is not so great with feedback to the list and I have
> already mentioned that in the past, EMC employees never reply as it
> seems that EMC do not encourage open, two-way discussion on this list
> (or elsewhere). Regardless, I will give it a try:
> 
>    . Avamar integration is poor. As already mentioned on this thread,
> EMC should have Avamar better integrated in Networker. I have given up
> hope of getting any useful dedup from EMC and waiting for ZFS dedup to
> solve this problem (AFTD on top of deduped ZFS).
> 
>    . AFTD concurrent operations was not improved since 7.2. You cannot
> run staging, cloning and recoveries at the same time and I still need to
> apologize to my users who are trying to run recoveries in the middle of
> the day while staging is running.
> 
>    . AFTD staging design is naive. Any non-trivial installation must
> script staging so that it will work in a more decent way. Some
> improvements that are missing:
> 
>    Staging in small chunks (I have seen a staging session that tries to
> stage over 1Tb in a single session). This allows for recoveries to run
> after a chunk finishes and also releases space on the AFTD faster (you
> don't have to wait till 1Tb gets to the tape before you can use the
> space). My script now tries to stage 15Gb at a time.
> 
>    Staging which can be triggered only when there is no (or little)
> activity on the AFTD.
> 
>    Delayed staging (copy to tape now, delete when space is actually needed).
> 
>    Finally fixing the bug from 7.2 which makes nsrstage wait ~2 seconds
> for every save set between the end of the clone and the start of the
> deletion of the savesets from the AFTD.
> 
>    . Dedup AFTD (like Data Domain).

>From what I have been told by people at EMC, some of the features on your
list are planned for upcoming NetWorker versions. EMC doesn't respond to
feedback on this list here, but some EMC people definitely do read the
postings here and give them due consideration, especially on an issue that's
reported by numerous people.

Two months ago, I saw a very detailed roadmap at the previous NetWorker
Users Group meeting in Philadelphia by Mickey Friedman. As I recall from
Mickey's presentation, extensive work is being done to improve how AFDTs
work. EMC acknowledged that the AFTD feature needs improvement. I don't
recall what was said about Avamar though.

I would also like to see more controls built into the AFTD staging feature.
Your idea of staging in small chunks is great. I would also like to see a
feature where I can stop the staging, then restart it. This would come in
handy if the tape to which the data is being staged begins to show i/o
errors during the staging process. It would be good to simply set the tape
device to service mode and have NetWorker immediately begin staging to a
different tape drive.

I can also tell you that EMC does some focus group work with making
improvements to NetWorker's user interface via the console manager. A lot of
features that appear in the latest two versions have come from feedback on
this list and focus group sessions that EMC holds at the annual EMC World
conference, which will be in Orlando this year.

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER