Backup Central

Download gawk and pipe the reulsts of bpdbjobs -mostcolumns to it.

I'llwrite the awk line for what you need when i get to work tonight, but basically gawk is going to do two things for you here - it will let you easily filter that output down to the jobs where the parent ID and the JobID match (just the parents) and the systime() and strftime() functions will let you convert between the epoch style dating and something more human readable.

It's ultimately just a long one line command that way.
________________________________
From: veritas-bu-bounces < at > mailman.eng.auburn.edu [veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of judy_hinchcliffe < at > administaff.com [judy_hinchcliffe < at > administaff.com]
Sent: Thursday, February 12, 2009 12:55 PM
To: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: [Veritas-bu] bpdbjos and just parent

NB 6.5.3

I am trying to use this command to get just the parent jobs
in mm/dd/yy date formant and not unix epoch format.

(I will settle for all jobs if need be)

So I want to supply a list of servers in a for loop
Get the parent job for that client
What I am looking for is the active start and end times

But I am having issues getting that in mm/dd/yy format.

Anybody got any ideas or other commands I could use?


_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

This will return the client, policy, schedule, start time, end time, and exit status of just parent jobs of type backup. If you list the servers you want in a text file, prefixed with a ^ you could pipe this output to grep and evaluate it against that file as a set of regular expressions and just get the lines that begin with the server names you want. (Probably sufficient to evaluate it as text if those server names appear anywhere on the line, but just in case.)

Line will probably wrap - but this is one line.
bpdbjobs -most_columns | gawk -F"," "{if($1 == $34 && $2 == 0) print $7,$5,$6,strftime(\"%m\\%d\\%y %H:%M:%S\",$9),strftime(\"%m\\%d\\%y %H:%M:%S\",$11),$4}"


The format string \"%m\\%d\\%y %H:%M:%S\" could be expressed more succinctly as \"%D %T\" but that isn't working with the version I have, while the example above is. Also, if you really just want the start and end date as you specified you can obviously drop the %H:%M:%S from the output.

Patrick Sweeney
(978) 787-4553
patrick.sweeney < at > axcelis.com
I.T. Systems/Networks

For issues requiring immediate attention please contact the Solution Center
IT Solution Center
(978) 787-7777
Beverly.Helpdesk < at > axcelis.com
Axcelis Technologies

Have you searched here and here?

-----Original Message-----
From: veritas-bu-bounces < at > mailman.eng.auburn.edu [mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Sweeney, Patrick
Sent: Thursday, February 12, 2009 2:07 PM
To: judy_hinchcliffe < at > administaff.com; VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] bpdbjos and just parent

Download gawk and pipe the reulsts of bpdbjobs -mostcolumns to it.

I'llwrite the awk line for what you need when i get to work tonight, but basically gawk is going to do two things for you here - it will let you easily filter that output down to the jobs where the parent ID and the JobID match (just the parents) and the systime() and strftime() functions will let you convert between the epoch style dating and something more human readable.

It's ultimately just a long one line command that way.
________________________________
From: veritas-bu-bounces < at > mailman.eng.auburn.edu [veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of judy_hinchcliffe < at > administaff.com [judy_hinchcliffe < at > administaff.com]
Sent: Thursday, February 12, 2009 12:55 PM
To: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: [Veritas-bu] bpdbjos and just parent

NB 6.5.3

I am trying to use this command to get just the parent jobs
in mm/dd/yy date formant and not unix epoch format.

(I will settle for all jobs if need be)

So I want to supply a list of servers in a for loop
Get the parent job for that client
What I am looking for is the active start and end times

But I am having issues getting that in mm/dd/yy format.

Anybody got any ideas or other commands I could use?


_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

i have new problem

When i right click to Media > Driver > one of two driver, and click Diagnostics - Every tet fail

Robotics Diagnostics all sucessfull.

Please help me, thank

Wait for the next patch. (6.5.4) it will have a tree view

-----Original Message-----
From: veritas-bu-bounces < at > mailman.eng.auburn.edu
[mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Sweeney,
Patrick
Sent: Thursday, February 12, 2009 9:07 PM
To: judy_hinchcliffe < at > administaff.com; VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] bpdbjos and just parent

Download gawk and pipe the reulsts of bpdbjobs -mostcolumns to it.

I'llwrite the awk line for what you need when i get to work tonight, but
basically gawk is going to do two things for you here - it will let you
easily filter that output down to the jobs where the parent ID and the JobID
match (just the parents) and the systime() and strftime() functions will
let you convert between the epoch style dating and something more human
readable.

It's ultimately just a long one line command that way.
________________________________
From: veritas-bu-bounces < at > mailman.eng.auburn.edu
[veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of
judy_hinchcliffe < at > administaff.com [judy_hinchcliffe < at > administaff.com]
Sent: Thursday, February 12, 2009 12:55 PM
To: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: [Veritas-bu] bpdbjos and just parent

NB 6.5.3

I am trying to use this command to get just the parent jobs
in mm/dd/yy date formant and not unix epoch format.

(I will settle for all jobs if need be)

So I want to supply a list of servers in a for loop
Get the parent job for that client
What I am looking for is the active start and end times

But I am having issues getting that in mm/dd/yy format.

Anybody got any ideas or other commands I could use?


_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu



_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

There is a critical step that’s missing from the IIS Web Server setup for LiveUpdate (and one that took me weeks to get figured out):

You need to add a new MIME type of .flg of type application/octet-stream

I don’t think you need that for Apache, but not sure.


From: veritas-bu-bounces < at > mailman.eng.auburn.edu [mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Renee Carlisle
Sent: Thursday, February 12, 2009 2:06 PM
To: Ed Wilts; Michitsch, John
Cc: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Netbackup Live Update.


OK, first rule…forget the documentation Symantec gives you for this process..unfortunately it is not great. Here is what we use to get http going for lu for either Windows or Unix. We have had no problem with it now that we have the process down..and have been able to patch application clusters…so not sure what other people have seen there. Hope this helps:

OK, we usually use http. The disk share method is for Windows only and Unix clients can’t get to the patches, which kind of defeats the purpose in my opinion. So, we stick with http…here is the process we use:

For unix live update server http access:

<![if !supportLists]>1. <![endif]>Pull down patches to lu server/dir, unzip all in same dir (no subdirs for each patch bundle)
<![if !supportLists]>2. <![endif]>turn on http on lu server
/etc/init.d/apache start (first copy /etc/apache/httpd-example.conf --> httpd.conf)
<![if !supportLists]>3. <![endif]>ln -s <liveupdate dir> /var/apache/htdocs/<liveupdate dir name> or where ever www root is, above is default, this setting can be determined from the apach.conf file in /etc/apache/apache.conf
<![if !supportLists]>4. <![endif]>install nblu on any
/usr/openv/netbackup/bin/install_nblu
when prompted type in:
url=http://<liveupdateserver</<liveupdatedir>
<![if !supportLists]>5. <![endif]>Create lu policies per normal documentation and go

For windows live update server http access:

<![if !supportLists]>1. <![endif]>Install IIS, need windows install cd, control panel/add remove programs/windows components/application services/IIS (must go into details at each level to unselect and select components you want for IIS and www)
<![if !supportLists]>2. <![endif]>Start www service
<![if !supportLists]>3. <![endif]>Launch IIS manager, highlight default web site, right click, new virtual dir.
<![if !supportLists]>4. <![endif]>Put all patches in wwwrootliveupdate (c:inetpubwwwrootcreate virtual dir in IIS for liveupdate
<![if !supportLists]>5. <![endif]>Unzip all in same dir (no subdirs)
<![if !supportLists]>6. <![endif]>SET LU_SERVER=http://<luserver>/liveupdate (in installdir/symantec/liveupdate/LU_install.bat, then run it, make sure you have 3 exe's in there with it, LUCheck, LU_Registration LUSetup, all 4 files on NBU windows install cd)

**NOTES: keep all dir names and servernames lower case for unix/windows

Renee Carlisle
ServerWare Coporation

From: veritas-bu-bounces < at > mailman.eng.auburn.edu [mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Ed Wilts
Sent: Thursday, February 12, 2009 12:10 PM
To: Michitsch, John
Cc: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Netbackup Live Update.


On Wed, Feb 11, 2009 at 12:56 PM, Michitsch, John <JMICHITS < at > gannett.com ([email]JMICHITS < at > gannett.com[/email])> wrote:
Is anyone using this to upgrade client agents. All things I've seen/read are negative about the product. We wasted many hours trying to get it to work on version 6.5 when it first came out. 6.5.3 is supposed to fix some issues, I don't want to go down the LU road again without the possibility of success! Anyone have it up and running successfully?


There are many caveats to getting this going properly and some are severe enough that we're not able to implement it for us for some of the clients. We had a lengthy discussion with the developers of the client deployment packages at the last Symantec Tech Day in December so they're aware of the restrictions.

One of the biggest gotchas for us is that it's pretty much unusable in a cluster environment. One of its checks is to see if the client is actually doing a backup before it tries to upgrade itself. If you have multiple names for your hosts - and this is the case for close to 100% of the cluster users - this check can fail and your backup and/or upgrade job will fail.

In general, the NetBackup support for cluster clients is pretty poor. Yeah, Symantec has hacked "support" in there but it's simply the case of not breaking in a cluster but that's about it. LiveUpdate is just another example of where Symantec's cluster "support" fails to solve the problems properly.
.../Ed

Ed Wilts, RHCE, BCFP, BCSD, SCSP, SCSE
ewilts < at > ewilts.org ([email]ewilts < at > ewilts.org[/email])

Matt…ooops you are right…we did that too it must have just been missed when we documented the process.

From: Clausen, Matt R[EQ] [mailto:Matthew.R.Clausen < at > Embarq.com]
Sent: Friday, February 13, 2009 10:32 AM
To: Renee Carlisle
Cc: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Netbackup Live Update.



There is a critical step that’s missing from the IIS Web Server setup for LiveUpdate (and one that took me weeks to get figured out):

You need to add a new MIME type of .flg of type application/octet-stream

I don’t think you need that for Apache, but not sure.


From: veritas-bu-bounces < at > mailman.eng.auburn.edu [mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Renee Carlisle
Sent: Thursday, February 12, 2009 2:06 PM
To: Ed Wilts; Michitsch, John
Cc: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Netbackup Live Update.


OK, first rule…forget the documentation Symantec gives you for this process..unfortunately it is not great. Here is what we use to get http going for lu for either Windows or Unix. We have had no problem with it now that we have the process down..and have been able to patch application clusters…so not sure what other people have seen there. Hope this helps:

OK, we usually use http. The disk share method is for Windows only and Unix clients can’t get to the patches, which kind of defeats the purpose in my opinion. So, we stick with http…here is the process we use:

For unix live update server http access:

<![if !supportLists]>1. <![endif]>Pull down patches to lu server/dir, unzip all in same dir (no subdirs for each patch bundle)

1. turn on http on lu server

/etc/init.d/apache start (first copy /etc/apache/httpd-example.conf --> httpd.conf)
<![if !supportLists]>3. <![endif]>ln -s <liveupdate dir> /var/apache/htdocs/<liveupdate dir name> or where ever www root is, above is default, this setting can be determined from the apach.conf file in /etc/apache/apache.conf

1. install nblu on any

/usr/openv/netbackup/bin/install_nblu
when prompted type in:
url=http://<liveupdateserver</<liveupdatedir>
<![if !supportLists]>5. <![endif]>Create lu policies per normal documentation and go

For windows live update server http access:

<![if !supportLists]>1. <![endif]>Install IIS, need windows install cd, control panel/add remove programs/windows components/application services/IIS (must go into details at each level to unselect and select components you want for IIS and www)

1. Start www service
2. Launch IIS manager, highlight default web site, right click, new virtual dir.
3. Put all patches in wwwrootliveupdate (c:inetpubwwwrootcreate virtual dir in IIS for liveupdate
4. Unzip all in same dir (no subdirs)
5. SET LU_SERVER=http://<luserver>/liveupdate (in installdir/symantec/liveupdate/LU_install.bat, then run it, make sure you have 3 exe's in there with it, LUCheck, LU_Registration LUSetup, all 4 files on NBU windows install cd)

**NOTES: keep all dir names and servernames lower case for unix/windows

Renee Carlisle
ServerWare Coporation

From: veritas-bu-bounces < at > mailman.eng.auburn.edu [mailto:veritas-bu-bounces < at > mailman.eng.auburn.edu] On Behalf Of Ed Wilts
Sent: Thursday, February 12, 2009 12:10 PM
To: Michitsch, John
Cc: VERITAS-BU < at > mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Netbackup Live Update.


On Wed, Feb 11, 2009 at 12:56 PM, Michitsch, John <JMICHITS < at > gannett.com ([email]JMICHITS < at > gannett.com[/email])> wrote:
Is anyone using this to upgrade client agents. All things I've seen/read are negative about the product. We wasted many hours trying to get it to work on version 6.5 when it first came out. 6.5.3 is supposed to fix some issues, I don't want to go down the LU road again without the possibility of success! Anyone have it up and running successfully?


There are many caveats to getting this going properly and some are severe enough that we're not able to implement it for us for some of the clients. We had a lengthy discussion with the developers of the client deployment packages at the last Symantec Tech Day in December so they're aware of the restrictions.

One of the biggest gotchas for us is that it's pretty much unusable in a cluster environment. One of its checks is to see if the client is actually doing a backup before it tries to upgrade itself. If you have multiple names for your hosts - and this is the case for close to 100% of the cluster users - this check can fail and your backup and/or upgrade job will fail.

In general, the NetBackup support for cluster clients is pretty poor. Yeah, Symantec has hacked "support" in there but it's simply the case of not breaking in a cluster but that's about it. LiveUpdate is just another example of where Symantec's cluster "support" fails to solve the problems properly.
.../Ed

Ed Wilts, RHCE, BCFP, BCSD, SCSP, SCSE
ewilts < at > ewilts.org ([email]ewilts < at > ewilts.org[/email])

On Thu, Feb 12, 2009 at 12:34 PM, Renee Carlisle <rcarlisle < at > serverwarecorp.com ([email]rcarlisle < at > serverwarecorp.com[/email])> wrote:


Curious to know what issues/what type of clusters you had trouble with? What method were you using? We generally use http and have updated clients in a application cluster without any issues.


All clusters, if they're configured as per most best practices, can fail to detect a running instance of NetBackup. If you have a physical server backup and a virtual server backup on the same physical server, neither know about each other. You could rip out the binaries in the middle of a running job.

This is the same issue on all active/passive clusters whether they're Windows or Unix. The only one that's immune is VMS because you don't need virtual servers.



Similarly, when you're dealing with virtual servers, you can't control how many instances of NetBackup are running on the same physical system. You can't even count how many client licenses you need from a master server since it can't tell how many unique physical servers there are.
.../Ed

Ed Wilts, RHCE, BCFP, BCSD, SCSP, SCSE
ewilts < at > ewilts.org ([email]ewilts < at > ewilts.org[/email])

After talking to Ed directly we confirmed a correction in 6.5.2 would address LiveUpdate in clusters as well as other installs with busy jobs. Since there has been some new information shared about LU on this thread, Symantec will pursue a technote documenting this information until manuals are updated.
Etrack Incident = ET1053371Description:     We now cancel a LiveUpdate if any of the following specific processes    are running on the machine:  bpbkar (bpbkar32), tar (tar32), bpbrm, bptm,    or bpdm.  This avoids the problem of the install process canceling    running jobs.
-----Original Message-----On Thu, Feb 12, 2009 at 11:09:51 AM, Wilts, Ed <ewilts at ewilts.org>
One of the biggest gotchas for us is that it's pretty much unusable in acluster environment.  One of its checks is to see if the client is actuallyDescription: 0Description: 1Description: 2

Bill_Coleman < at > symantec.com
Office: 651 746 7118

With Sun Server T5140, Solaris 10, Netbackup 6.5, SL48
I want use MpxIO ( must use ) how can i config netbackup driver detect all robo and driver

Dear KinAn,
the best practise is you should use different HBA port to connecting
host to disk and tape library, After you enable mpxio then you need to
disable only the port connected to tape library at fp.conf file.

hope this helps..

Regards,
hmw

KinAn wrote:
With Sun Server T5140, Solaris 10, Netbackup 6.5, SL48
I want use MpxIO ( must use ) how can i config netbackup driver detect all robo and driver

+----------------------------------------------------------------------
|This was sent by kinan_kazuki < at > yahoo.com via Backup Central.
|Forward SPAM to abuse < at > backupcentral.com.
+----------------------------------------------------------------------


_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

But, I connect every over SAN. How to config like you talk? ( Mr Hermawan)

Backup server ( port1), 6140 ( Controller A ), SL48 ( Driver 1) connect to SAN SW 1

Backup server ( port2), 6140 ( Controller B ), SL48 ( Driver 2) connect to SAN SW 1

Please help me. Backup server still free 1 port on HBA

I have other question

1. Server , Tape, Storage connect to SAN SW ( SAN )
2. Server connect to Tape ( Direct )
Server connect to Storage ( Direct )

Have any diffirent between 1 and 2 when backup data from Storage to Tape. What's faster than?

KinAn,
base on my experience, since you need to enable mpxio for disk redundancy, you can't use the same port to connect to SL48, because SL48 infrastucture has daisychain between robotic and tape drive, this mean mpxio must be disable.
if your SL48 has 2 tape drives, you can try to define which tape drive has daisychain to the robotic then use the another free hba port to connect to this drive...
again.. dont forget to disable mpxio at fc.conf file only to hba port use for tape device.

hope this help.

Regards,
hmw

----- Original Message -----
From: KinAn <netbackup-forum < at > backupcentral.com>
Date: Wednesday, August 19, 2009 10:47 am
Subject: [Veritas-bu] Solaris 10 run Netbackup cannot detect Sun StorageTekSL48??


But, I connect every over SAN. How to config like you talk? ( Mr
Hermawan)
Backup server ( port1), 6140 ( Controller A ), SL48 ( Driver 1)
connect to SAN SW 1

Backup server ( port2), 6140 ( Controller B ), SL48 ( Driver 2)
connect to SAN SW 1

Please help me. Backup server still free 1 port on HBA

+-------------------------------------------------------------------
---
|This was sent by kinan_kazuki < at > yahoo.com via Backup Central.
|Forward SPAM to abuse < at > backupcentral.com.
+-------------------------------------------------------------------
---


_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist - Veritas-bu < at > mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu