Back Discrepancy
DrumDude
for the 3rd weekend in a row.
It was always fast now it is very slow.
OES2SP1/SLES10SP2
Here is the current job that is still running.
Directories: 121906
Files: 86354258 Elapsed time: 44:44:00
Bytes: 1248091.5 MB
I read this as 1.2TB: but look at the size and used space is only 790GB.
df -h
SANVOL1 2.0T 790G 1.3T 39% /media/nss/SANVOL1
/etc/fstab
SANVOL1 /media/nss/SANVOL1 nssvol
noauto,rw,name=SANVOL1,ns=long,norename,noatime,nodiratime 0 0
State: mounted
│
│││ Supported Name Space(s): Dos Long Macintosh Unix
│
│││ Name Space: long
Can anyone see what I am missing?
DrumDude
It keeps backing up the same files in a folder over and over again.
There is a problem with files in a ONE folder the backups get caught in
an infinite loop.
Any ideas?
Bob-O-Rama wrote:
> DrumDude;1921980 Wrote:
>> I have been having a problem with Full Backup of Symantec Backus Exec
>> for the 3rd weekend in a row.
>> It was always fast now it is very slow.
>> OES2SP1/SLES10SP2
>> Here is the current job that is still running.
>> Directories: 121906
>> Files: 86354258 Elapsed time: 44:44:00
>> Bytes: 1248091.5 MB
>> I read this as 1.2TB: but look at the size and used space is only
>> 790GB.
>>
>> df -h
>> SANVOL1 2.0T 790G 1.3T 39% /media/nss/SANVOL1
>>
>> /etc/fstab
>> SANVOL1 /media/nss/SANVOL1 nssvol
>> noauto,rw,name=SANVOL1,ns=long,norename,noatime,nodiratime 0 0
>>
>> Can anyone see what I am missing?
>
> How many files do you actually have? Its showing 86354258 ( 86 million
> ) backed up so far - which seems very high. -Its only a problem if you
> don't have 86 million files. - :-)
>
> You should be able to see what folder / files your BE engine is backing
> up right now, and if the backup is progressing BE may just be stuck -
> or its backing up a ginormous sparse file, or perhaps a hard link going
> in circles.
>
> I also just find BE does stupid stuff on occasion and the reboot of the
> engine is necessary to unclog it. But best to first see what it was
> backing up. 44 hours for backup points to the engine being stuck or
> something odd in the file system.
>
> So questions to answer: what is BE doing right now? Note the folder
> and file and if the backup is progressing. Second, does the BE view of
> the volume match reality - are there really 86 million tiny files?
> Then go where that leads you.
>
> -- Bob
>
>
DrumDude
problem resides. It keeps backing up the same files over and over again.
Weird?
I am going to exclude the folder for now although I would like to
understand what the real problem is for preventative measures.
Any ideas?
Thanks
DrumDude
> DrumDude;1922832 Wrote:
>> As a test I tried backing up that folder as a single job. The same
>> problem resides. It keeps backing up the same files over and over
>> again.
>> Weird?
>> I am going to exclude the folder for now although I would like to
>> understand what the real problem is for preventative measures.
>>
>> Any ideas?
>>
>
> You'll need to watch BE as it backs up the files, or better, enable the
> verbose backup logs which will show individual files being backed up...
> and see exactly where the loop begins.
>
> Sorry if it sounds like the movie Groundhogs Day with you cast in place
> of Bill Murray.
>
> -- Bob
>
>
Crazy this problem is random to any folders.
I think it may have something to do with noatime,nodiratime on the mount
point that someone suggested for better performance in these threads and
the NSSmanual.
Can anyone confirm or have knowledge with that?
>
A.25 noatime and atime Commands
The /noatime and /atime commands for a volume allow the administrator to
control whether
access times are updated when files and directories are read. They are
available for NSS on Linux
and NetWare beginning with OES 2 SP1 and NetWare 6.5 SP8. The setting
persists across reboots.
Using /noatime is useful for backup, news servers, and mail servers
where the extra disk activity
associated with updating the access time is not desired. Avoiding the
writes associated with updating
the access time can result in measurable performance gains.
DrumDude
Upon further investigation if I browse the files of a problematic backup
folder via a drive mapped or from the server console source
/media/nss/VOLUMES.....
I see 52 objects/files an exact match.
If I browse from the backup agent it is infinite.In other words after
the 52nd file browse, it starts at file 1 again????
Just whacked!!!
The only change is OES2SP2/SLES10SP3.
Any ideas?
Thanks