Hi
This morning it is asking for a tape again.
It seems it always is an Exchange saveset which is the culprit. (not sure
though)
The original saveset is on Diskbackup.001 (0, c) Savetime 31.01.2012 23:16:32
(Pool Diskbackup) 555GB
There is a cloned saveset on two tape volumes
LBL0116L3 (1000.0, h) 493 GB , LBL0088L3(100.0, t) 64 GB Savetime 31.01.2012
23:16:32 (Pool Tapebackup)
It starts to stage from the Diskbackup AFTD and for some reason in the middle
of the stage process it wants to clone the tail saveset on LBL0088L3 which has
already been staged to the Tapebackup pool.
I am clueless here.......
..and the script I was running is just for troubleshooting, it is not an option
to be used as a solution.
E;>
----------------------
From: DE VENTE Gert
Sent: 1. februar 2012 08:32
Subject: nsrstage
set SSIDFILE=SSIDs_to_be_cloned.txt
mminfo -r ssid -q pool=Diskbackup > %SSIDFILE% for /f %%I in ('type
"%SSIDFILE%"') do (
nsrstage -vvvvv -b Tapebackup -m -S %%I
)
Try
mminfo -r ssid -q "family=disk,copies<3"
This will handle only savesets which have only 2 copies on the ATFD (RW + RO).
Your selection does not check for this and as the cloneID for tape-clone is
less than the RW or RO, a saveset selected from tape will be staged based on
the SSID.
With kind regards,
ing. Gert de Vente
-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU] On
Behalf Of Eivind Antonsen
Sent: 31. januar 2012 12:46
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] staging problem
The script has now reached a SS which makes the stage process ask for a tape.
When searching for the SS under "Media->Save Sets" I find that the SS
"original" is 555GB and is located on the Diskbackup device.
Diskbackup.001 (0, c)
It has two clones, one from the stage process last night and one from my script
now running
BTK301L3 (1000.0, h), LB0116L3 (8.0, t)
BTK292L3 (1000.0, h), LB0088L3 (100.0, t)
The stage process is asking for tape LB0088L3
When I list the savesets on the tape I find the one it wants to read and it is
33GB in size and has a "tb" flag.
I know if I run a "nsrmm -d -S SSID" and delete the saveset on LB0088L3 and
then start the stage process again it will not ask for a tape when performing
stage.
I belive the (0, c) on the diskbackup device means it is complete..?
What makes it want to read from a tail of a cloned saveset ?
E;>
-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU] On
Behalf Of Eivind Antonsen
Sent: 31. januar 2012 12:06
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] staging problem
This morning it was asking for a tape again, I terminated the nsrstage process
and I am currently running the following script to see if it will ask for a
tape.
set SSIDFILE=SSIDs_to_be_cloned.txt
mminfo -r ssid -q pool=Diskbackup > %SSIDFILE% for /f %%I in ('type
"%SSIDFILE%"') do (
nsrstage -vvvvv -b Tapebackup -m -S %%I
)
-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU] On
Behalf Of Eivind Antonsen
Sent: 30. januar 2012 09:25
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] staging problem
Hello
The raid controller issue is resolved, but the problem persists.
I've been trying to figure this one out, but I'm really stuck.
NW server Win 7.6.3
One AFTD device (Media pool "Diskbackup") One Library (LTO-3) with one drive
(Media pool "Tapebackup")
Staging policy (have tried different settings, but this is current)
Source: AFTD "Diskbackup"
Target: "Tapebackup"
High watermark 50%
Low watermark 10%
Max storage period 4 days
Recover space interval 15 minutes
File system check 20 minutes
Ok, so heres the problem:
Backup writes to "Diskbackup" media pool on the AFTD The AFTD reaches watermark
and starts to stage.
When the staging starts it requests to read from AFTD "Diskbackup" and (!!) a
tape in the "Tapebackup" media pool.
The "Tapebackup" tapes should only contain already staged savesets.
The mediapool "Tapebackup" is not target for anything except the staging policy.
So, if the AFTD ran out of space it would request a "default" media and not
write to the "Tapebackup" tapes, or am I wrong ?
Why would the stage process need to stage already staged savesets ?
I've checked the savesets on the requested tapes and sometimes they contain
head,middle or tail savesets, and sometimes not.
I've tried to set "recover space" interval down to 1 minute, because I've seen
that NW sometimes only "recover space" from 30 Savesets at a time.
Because there is only one drive in the library this problem causes the server
to halt all backups, waiting for space on the AFTD and for read from the
mentioned tape.
This in turn causes the Exchange server to halt when the logs fill up, among
other things.
Any suggestions ?
Thanks
Eivind
-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU] On
Behalf Of Eivind Antonsen
Sent: 2. desember 2011 10:24
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] staging problem
Hi
Thank you for your reply.
I don't see how it can write the tail of a backup to a tape when there is only
one drive and it is used to receive the stage job.
None of the savesets have other flag than 'c'
But, there is an issue with the raid controller on the backupserver and I will
wait for it to be resolved to see if it may somehow be the casue.
Thanks
Eivind
-----Original Message-----
From: Small, Joshua [mailto:joshua.small AT citi DOT com]
Sent: 24. november 2011 01:06
To: 'NETWORKER AT LISTSERV.TEMPLE DOT EDU'; Eivind Antonsen
Subject: RE: staging problem
There was some changes in 7.6 that changed the way AFTD volumes work.
Previously, if your AFTD filled up during a save, the save would wait for space
to become available in that volume.
Now, AFTDs are handled like a tape volume, and will continue the saveset on
another volume.
Maybe your AFTD ran out of space, and the tail of the saveset was written to a
tape?
So when you stage that ssid, it would need to read the start of the saveset
from AFDT, and the end of it from tape - because staging acts on savesets, not
chunks of savesets in a volume.
You could identify any continued savesets with 'mminfo -av', the 'c' flag in
the output is complete, h is head (start of saveset), m is middle and t is tail.
Those last three flags all indicate that more than one volume is needed for
that saveset.
Or, something like this will grab a list of all ssids on your AFTD volume, and
then check what volume types and names contain each saveset
$ mminfo -av -q 'volume=your_aftd_volume_name' -r ssid | perl -ne
'chomp;system("mminfo -q ssid=$_ -xc, -r ssid,type,volume|grep -v ssid");'
(but won't work in a windows environment unless you've got perl/grep available,
there's probably ways to do the same thing from windows with different tools
though)
Regards,
Josh
> 7.6.2.681 (server and client)
> The staging starts and requests to read from AFTD and a tape(!!)
> What could make it need to read from a tape ?
> How can I locate which savesets on this tape that it is trying to stage ?
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
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
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
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
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
|
