Conference decwet::networker

do you have:

1) Clones as SS
2) HW Links 
[Posted by WWW Notes gateway]

1) no clones as SS

2) what is the point with the HW links?

it could be doing a full... post 

 	savegroup -pnv -c client-name group-name-containing-client

verbatim so we can see too.

hessen.frs.dec.com:/                      level=7
hessen.frs.dec.com:/usr                   level=7
hessen.frs.dec.com:/var                   level=7
 5/22/97  7:13:28 savegroup: Run up to 4 clients in parallel
 5/22/97  7:13:28 savegroup: hessen.frs.dec.com:probe                   started
savefs -s tbla2.frs.dec.com -g systeme -p -n -l full -R -v -F / /usr /var 
 5/22/97  7:13:34 savegroup: hessen.frs.dec.com:probe succeeded.
rcmd hessen.frs.dec.com, user root: `savefs -s tbla2.frs.dec.com -g systeme -p
-n -l full -R -v -F / /usr /var'
asavegrp: authtype nsrexec
type: NSR client description;
pools supported: Yes;
arch: decaxp;
CPU type: alpha;
CPUs: 1;
hosts: 1;
IP address: 16.185.224.20;
kernel arch: OSF1;
License units: 1050;
NAS Client: No;
OS: OSF1 V4.0;
version: V4.2A;
save set: path=/, arg=/, level=full, diskno=2, max_sessions=1, stype=save,\
path=/usr, arg=/usr, level=full, diskno=1, max_sessions=1, stype=save,\
path=/var, arg=/var, level=full, diskno=0, max_sessions=1, stype=save ;
parallelism: 8
--- Probe Summary ---

hessen.frs.dec.com:/                  level=7, dn=2, mx=1, vers=pools, p=8
hessen.frs.dec.com:/      level=7, save as of Mon Mar 17 22:12:08 GMT+0100 1997
hessen.frs.dec.com:/usr               level=7, dn=1, mx=1, vers=pools, p=8
hessen.frs.dec.com:/usr   level=7, save as of Mon Mar 17 22:06:16 GMT+0100 1997
hessen.frs.dec.com:/var               level=7, dn=0, mx=1, vers=pools, p=8
hessen.frs.dec.com:/var   level=7, save as of Mon Mar 17 22:10:54 GMT+0100 1997
hessen.frs.dec.com:index             level=7, dn=-1, mx=0, vers=pools, p=8
hessen.frs.dec.com:index  level=7, save as of Thu May  1 22:47:57 GMT+0200 1997

11860           22053  637 MB  3/17/97  full
11897           19234  314 MB  3/18/97    incr
11943           19136  313 MB  3/19/97    incr
11983           19198  313 MB  3/20/97    incr
12137           19621  316 MB  3/24/97    incr
12180           19336  318 MB  3/25/97    incr
12219           19341  314 MB  3/26/97    incr
12371           19578  329 MB  4/11/97    incr
12511           19400  317 MB  4/14/97    incr
12560           19995  352 MB  4/15/97    7
12604           19511  317 MB  4/16/97      incr
12648           19957  370 MB  4/17/97      incr
12819           20062  322 MB  4/22/97      incr
12862           20247  410 MB  4/22/97    7
12909           19681  320 MB  4/23/97      incr
12952           19249  314 MB  4/24/97      incr
13038           19229  315 MB  4/28/97      incr
13070           19202  314 MB  4/28/97      incr
13113           19908  411 MB  4/29/97    7
13361           19631  374 MB  5/05/97      incr
13401           19086  307 MB  5/06/97      incr
13446           19485  309 MB  5/07/97      incr
13490           20537  471 MB  5/08/97    7
13537           19204  307 MB  5/13/97      incr
13582           19256  308 MB  5/14/97      incr
13627           20493  473 MB  5/15/97    7
13841           19342  313 MB  5/20/97      incr
13882           19509  308 MB  5/21/97      incr

You need to see what is being backed up in an incremental
on this client. Maybe you have a large file which is touched
daily.

There's a NetWorker command called nsrinfo which allows you
to see what is in a client's online index for a particular
save, among other things. Here's how to do it (from the nsrinfo
man page example, tried on my workstation):

root@cochiti[20]> mminfo -r nsavetime -v -N / -c cochiti -ot |tail -1   
  864314936
root@cochiti[21]> nsrinfo -t 864314936 cochiti.zso.dec.com 

Result is a list of all files backed up for that save. My example 
looks at the last save; you will want to make sure you specify one 
of the large incremetals you're interested in, as well as the 
appropriate filesystem in the mminfo command.

When I tried it, the fully qualified client name was required
by nsrinfo.

Debbie

Debby,

today I did two incremental backups on my workstation by manually starting
them
from nwadmin's "customize/Groups..." dialogbox. The first incremental totalled
to 18178 files and 267MB, the second had 17473 files and 266MB. Then I
generated
a list of the saveset contents using the commands you proposed. 
Result: the 17473 files from the second saveset had been backed up twice
within
half an hour.
How does Networker determine what files to include into an incremental
backup??

Uwe
 
[Posted by WWW Notes gateway]

save -l incr saves all files changed (mtime) since last save. if mtime gets
 changed for *any* reason, files are saved by NSR.

save -l <number> saves files that have mtime changed since last level "x"
 save.

schedule determines the level for a savegroup - check your schedule.

regardless of what level is specified, if NetWorker determines there are no
 entries in the index, it does a full backup. This can occur anytime should
 the name of the index not match *precisely* the name of the client being
 saved (for one example).

Uwe,

   If the files did not have their mtime values changed between the 
two incrementals you ran, maybe you have a directive in effect for
them. The "always" directive causes a file or directory to be backed
up regardless of its last change time. Look at the nsr_directive and
uasm man pages for more information.

Debbie

Debbie,

I didn't find any nsr directives in effect. Also, when I looked for modified
files using

	find /usr -mtime -2

I found about 1000 files that had been changed since Monday. Doing a

	savegrp -n -C Quarterly Test
or
	savegrp -n -l inc Test

showed the following output:

  hessen.frs.dec.com:/ 685 records 127 KB header 13 MB data
  hessen.frs.dec.com:/ 13 MB estimated
  hessen.frs.dec.com:/var 323 records 62 KB header 5.0 MB data
  hessen.frs.dec.com:/var 5.1 MB estimated
  hessen.frs.dec.com:/usr 17693 records 3.5 MB header 279 MB data
  hessen.frs.dec.com:/usr 282 MB estimated
  tbla2.frs.dec.com:/var/nsr/index/hessen.frs.dec.com 2 records 1 KB header
6.5 MB data
  tbla2.frs.dec.com:/var/nsr/index/hessen.frs.dec.com 6.5 MB estimated

I really don't know what to do now. Do you have any other ideas?

Uwe
[Posted by WWW Notes gateway]

Uwe,

  Do the 17693 files have anything in common? Maybe there is something 
about the data that can help solve the mystery. Other than that, I
don't know what else to suggest.

Debbie

a customer called in an IPMT to NetWorker Engineering swearing NetWorker
was broken. After 3 months of arduously researching this, it came down
 to a cron entry that touched (altered the m-time) of every file in the
 system, thus making NetWorker do a full (apparently) backup.

moral: watch your cron jobs.

I finally found the reason for Networker's strange backup behaviour!

When I installed DU 4.0B on my workstation the system date had been set to
Jun 02 1997. I didn't realise that at first and it did no harm to my daily
work.
Then I installed the Networker client and experienced the large incremental
backups. Looking for reasons I found several files and directories dated
Jun 02 97. So I touched each of them to reset the modification date and time
to the current values. However, this didn't change Networker's backup
behaviour.

Not knowing what else to do I entered the note here to ask for help.

Today (Jun 04 1997) I looked again across the saveset indexes of my
workstation
and found that the size of the last incremental backup had shrinked down to
about 3 to 4 MB!

So I guess that touching the files in question was not sufficient. I think I
should have done another full backup after touching the ill-dated files to
"reset" Networker as well.

Sorry for causing any headaches and many thanks for all your answers!!

Kind regards

Uwe.
[Posted by WWW Notes gateway]