Hi to everyone
Our environement is Solaris Sparc 64bit with Netbackup 7.0.1 on SunOS 5.0 (Solaris 10) with a Windows 2003 x86 Standard Proxy as Man in the Middle to a VNC/CIFS-Disk-Platform
There exists two Policies each with two Shedules :
1th Full -Backup
2nd Cummulative Incr Backup
Exclude-Lists are created with the following content :
CIFS_Gem_T1 : *
CIFS_Gem_T2 : *
Include-Lists are created with the following contends :
Based on the description in the documentation, the wildcard parameters and the syntax rules for Exclued and Include lists should be correct. But every Test with the Policy CIFS_Gem_T2 leads to a backup beginning with the alphabetic character b .
Is there an inappropriate use of parameters and/or rules ?
is there a missing patch or a damaged file on the system ?
Many thanks for any help
It's attributed to the fact, that we use (at the moment) UFS as File-System and the amount of Data is over 2 TB.
So we tried to divide the Backup2Disk untill the moment we can move to ZFS.
The Backup-Targets are a List of 100 to n of public Shares of a File Server, whích are "breathing" on the requirements of the departements.
btw. a Test with deactvating Multistreaming did not change anything as mentioned in another discussion.
Well, what I want to say is..
Why don't you directly write "\\servername\share\[l-zL-Z]*" into Backup Selections in CIFS_Gem_T2 policy instead? I think there are no need to use exclude/include list to back it up. Why do you bother to exclude everything(* in exclude list) and include files again?
Sorry, Good Idea, But not the end of my problem :
04.04.2013 12:15:52 - requesting resource sbtit1-n-3
04.04.2013 12:15:52 - requesting resource anthony.NBU_CLIENT.MAXJOBS.bu-strfile1-nbu
04.04.2013 12:15:52 - requesting resource anthony.NBU_POLICY.MAXJOBS.CIFS_Gem_T2
04.04.2013 12:15:52 - granted resource anthony.NBU_CLIENT.MAXJOBS.bu-strfile1-nbu
04.04.2013 12:15:52 - granted resource anthony.NBU_POLICY.MAXJOBS.CIFS_Gem_T2
04.04.2013 12:15:52 - granted resource MediaID=@aaaaB;Path=/dssu/backupvol15/sbtit1-n-3;MediaServer=anthony
04.04.2013 12:15:52 - granted resource sbtit1-n-3
04.04.2013 12:15:52 - estimated 0 kbytes needed
04.04.2013 12:15:53 - started process bpbrm (pid=17069)
04.04.2013 12:15:55 - connecting
04.04.2013 12:15:56 - connected; connect time: 0:00:00
04.04.2013 12:16:03 - Error bptm (pid=17083) system call failed - Connection reset by peer (at child.c.1296)
04.04.2013 12:16:03 - Info bpbrm (pid=17069) from client bu-strfile1-nbu: TRV - object not found for file system backup: \\strfile1\admin$\Gemeinsam\[l-zL-Z]*
04.04.2013 12:16:03 - Error bptm (pid=17083) unable to perform read from client socket, connection may have been broken
04.04.2013 12:16:03 - Error bptm (pid=17079) media manager terminated by parent process
04.04.2013 12:16:06 - Error bpbrm (pid=17069) could not send server status message
04.04.2013 12:16:02 - begin writing
04.04.2013 12:16:07 - end writing; write time: 0:00:05
none of the files in the file list exist (71)
I act on your proposal. But it did not solve the problem. the backup fails with the following error-code :
(71) none of the files in the file list exist"
The trouble ist, a second try with a full exclude/include-list for all schedules-names also fails.
For Windows clients, the only wildcards supported in the filelist are * and ? - so Yasuhisa's suggestion would be a valid one for unix clients, but not here.
That being said, I don't see a problem with your include/exclude files.
Have you verified that NBU on the client isn't running under the SYSTEM account (http://www.symantec.com/business/support/index?page=content&id=TECH6099)? I'd guess not since you got some data backed up, but just so we know...
In your policies, what are the backup selections set to?
Oh, i have forgot that! Sorry!!
If subdirectory starting with 'b' isreally backed up by CIFS_Gem_T2(really?), you should log a call with Symantec. Or sort it by expanding blacket by hand and write into backup selection(30 lines for CIFS_Gem_T2).
BTW, i know similar issue with NetBackup 6.x on RHEL. Thus caused by unexpected locate redifinition in inetd - nothing wrong in NetBackup. In some locale, blacket is treated with abnormal character order.
I'm not sure system locale affects character order in Windows, but would be better to tell your system locale setting to symantec if you call.
Gosh! Might it be, that Netbackup scans the whole file-/foldername ? Never mind, at which position of the word, the alphabetic character will be placed ?
That makes no sense at all; but could clarify the behaviour.
@bills : NBU is running with a special account with administrative rights
The issue I wrote above is caused by inetd on RHEL so this is not appicable as it is on Windows. Only once I saw this several years ago, and I can not remind the detail.
Now, I tested similar scenario with NetBackup 126.96.36.199, Windows 2008 R2, United States(region and locale), and local disk. But this issue does not occur in my test environment. Folders are picked up as expected.
As i wrote before, if you want to sort this quickly, use Backup Selections list expanded by hand.
Otherwise, it is time to open a support case.
I don't see anything wrong with what you posted - it should perform as you expect.
Does the T1 policy back anything up?
In your tests, did you start out with a full then do the incrementals? Technically an incremental against a directory that had never been backed up will result in a full backup, but it seems to be a grey area.
It could also be that the $ in the exclude list is confusing the regular expression? You could try replacing admin$ with * and see what happens...
If you turn on verbose debugging, I think one of the logs goes into details on what is being excluded/included and why - you might try that to see if you can figure out what's going on.
Technically yes, but the limitations of the UFS-Filesystem hinders us to make a full backup, because T1 backups anything too.
The test were carried out as Cummulative/Incremental Backups.
Your proposal with the $ did not confuse the system. Neither the expression /*/ nor the expression /admin?/ work. Both faiiled with the Status 
To workaround the problem we consolidated the policies in a uniform manner, that for all alphabetic and numeric charakters as a line within.
To optimize the use ,
[x] Allow multiple Data streams
[x] Limit Jobs per Policy
are used, to group the upcoming jobs.
Different Tests gave us the technical confirmation, that both Policys could be started parallel with up to four streams and 20 MB/s on every Side.
p.s. At this moment, we hold an support call for this problem at symantec support center. At least for the meantime we use the outlined above method.
I will mark your proposal as solution because any other tests on our own -plus the assistance of bill -and and an open call at symantec adduce no clue to search any longer.
Thanx to either help
The wildcard-mechanism gives me the willies . I trie to save files and folders beginning from a special place.
But; the using of \[alphanumeric character]* -Method leads me to two Questions
1. will Netbackup will only backup files which contains this [alphanumeric characters] ?
In other Words, will it deny the others ?
2. Which method can be used to tell Netbackup to backup recursive ; ?
or in other Words , will the Method [a-kA-K]* not leed to the same phenomenon ?
As a Resume : Is there a Minimum-Requirement to Netbackup needed to sail around this cliffs.
As a backup selection this will backup anything *directly* within \path that starts within the range a-k or A-K - if these happen to be folders it will backup *anything* within these folders beit Ziggy or Aardark or 10greenbottles.