Veritas-bu

[Veritas-bu] Configurable wait between retries? Was: Thoughts on rerunning a single failed stream?

2002-03-05 19:00:29
Subject: [Veritas-bu] Configurable wait between retries? Was: Thoughts on rerunning a single failed stream?
From: Fabbro.Andrew AT cnf DOT com (Fabbro, Andrew P)
Date: Tue, 5 Mar 2002 16:00:29 -0800
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C1C4A1.ED3FE190
Content-Type: text/plain;
        charset="iso-8859-1"

You know, I have our environment set to 4 tries per 6 hours (the setting I
inherited).  However, I find that if a job fails, it will make all its
retries very quickly - within a few minutes.  I see this consistently.  The
wakeup period is set to 10 minutes.
 
Is there a configurable "wait this long between retries" setting?
 

-----Original Message-----
From: Cornely, David [mailto:David_Cornely AT intuit DOT com]
Sent: Tuesday, March 05, 2002 2:19 PM
To: 'Veritas Netbackup '
Subject: RE: [Veritas-bu] Thoughts on rerunning a single failed stream?


Increase the # of tries per job to 2 or 3 within a 12 hour time frame.  This
way when 1 stream fails it will re-queue instead of all streams from that
class.  Also, if it becomes necessary to re-run it manually why not just
create a Special_Requests class with only the stream that needs to be backed
up?  We do this and it works just fine -- one thing to note is if you're
using vaulting be sure to give the schedule the exact same name and
retention so vaulting will catch it.
 
If you run bpbackup from the client (user-directed) you can use the -f
option to specify a list of files to back up.  If you run it on the master
server you cannot because it is the equivalent of running a manual backup
from the admin gui.  In that case your includes file for the class becomes
the effective file list to back up.
 

-----Original Message-----
From: Fabbro, Andrew P [mailto:Fabbro.Andrew AT cnf DOT com]
Sent: Tuesday, March 05, 2002 12:47 PM
To: 'Veritas Netbackup '
Subject: [Veritas-bu] Thoughts on rerunning a single failed stream?



We have some backup classes that have multiple streams.  When one stream
fails, the only way to rerun just that stream is to initiate a backup on the
whole client and then kill off the other streams.  This is a tedious process
because you have to lookup the file list on the stream that failed, compare
it to the file list for the streams that are running, kill them, etc.

Does anyone have a better solution?  In most cases, logging into the client
and using bpbackup is not an option. 

I'm almost tempted to break all of my classes into multiple classes, one
stream each, but that just seems ugly (or have other people tried that?).
Just rerunning the whole backup is not good because it chews up way too much
tape on, say, a 500Gb full ;)  

I've been playing with bpbackup on the master server to see if there is some
way of saying "backup these files on this client," but bpbackup apparently
doesn't support this - it either runs a schedule for a client or you use it
on the client.  There appears to be no way to say "here's a list of files on
client X, now go back them up as if we're running a full/incremental/etc."

Anyone have any ideas? 

-- 
Drew Fabbro 
fabbro.andrew AT cnf DOT com 
503-450-3374 (desk)  503-701-0369 (cell) 
"Everyone's a superhero.  Everyone's a Captain Kirk." 
   -- Nina Hagen 


------_=_NextPart_001_01C1C4A1.ED3FE190
Content-Type: text/html;
        charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<TITLE>Thoughts on rerunning a single failed stream?</TITLE>

<META content="MSHTML 5.50.4912.300" name=GENERATOR></HEAD>
<BODY>
<DIV><SPAN class=859035523-05032002><FONT face=Arial color=#0000ff size=2>You 
know, I have our environment set to 4 tries per 6 hours (the setting I 
inherited).&nbsp; However, I find that if a job fails, it will make all its 
retries very quickly - within a few minutes.&nbsp; I see this 
consistently.&nbsp; The wakeup period is set to 10 minutes.</FONT></SPAN></DIV>
<DIV><SPAN class=859035523-05032002><FONT face=Arial color=#0000ff 
size=2></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=859035523-05032002><FONT face=Arial color=#0000ff size=2>Is 
there a configurable "wait this long between retries" 
setting?</FONT></SPAN></DIV>
<DIV><SPAN class=859035523-05032002><FONT face=Arial color=#0000ff 
size=2></FONT></SPAN>&nbsp;</DIV>
<BLOCKQUOTE dir=ltr style="MARGIN-RIGHT: 0px">
  <DIV class=OutlookMessageHeader dir=ltr align=left><FONT face=Tahoma 
  size=2>-----Original Message-----<BR><B>From:</B> Cornely, David 
  [mailto:David_Cornely AT intuit DOT com]<BR><B>Sent:</B> Tuesday, March 05, 
2002 2:19 
  PM<BR><B>To:</B> 'Veritas Netbackup '<BR><B>Subject:</B> RE: [Veritas-bu] 
  Thoughts on rerunning a single failed stream?<BR><BR></FONT></DIV>
  <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
  class=639121522-05032002>Increase the # of tries per job to 2 or 3 within a 
12 
  hour time frame.&nbsp; This way when 1 stream fails it will re-queue instead 
  of all streams from that class.&nbsp; Also, if it becomes necessary to re-run 
  it manually why not just create a Special_Requests class with only the stream 
  that needs to be backed up?&nbsp; We do this and it works just fine -- one 
  thing to note is if you're using vaulting be sure to give the schedule the 
  exact same name and retention so vaulting will catch it.</SPAN></FONT></DIV>
  <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
  class=639121522-05032002></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT face=Arial color=#0000ff size=2><SPAN class=639121522-05032002>If 
  you run bpbackup from the client&nbsp;(user-directed) you can use the -f 
  option to specify a list of files to back up.&nbsp; If you run it on the 
  master server you cannot&nbsp;because it is the equivalent of running a 
manual 
  backup from the admin gui.&nbsp; In that case your includes file for the 
class 
  becomes the effective file list to back up.</SPAN></FONT></DIV>
  <DIV>&nbsp;</DIV>
  <BLOCKQUOTE style="MARGIN-RIGHT: 0px">
    <DIV class=OutlookMessageHeader dir=ltr align=left><FONT face=Tahoma 
    size=2>-----Original Message-----<BR><B>From:</B> Fabbro, Andrew P 
    [mailto:Fabbro.Andrew AT cnf DOT com]<BR><B>Sent:</B> Tuesday, March 05, 
2002 12:47 
    PM<BR><B>To:</B> 'Veritas Netbackup '<BR><B>Subject:</B> [Veritas-bu] 
    Thoughts on rerunning a single failed stream?<BR><BR></DIV></FONT>
    <P><FONT size=2>We have some backup classes that have multiple 
    streams.&nbsp; When one stream fails, the only way to rerun just that 
stream 
    is to initiate a backup on the whole client and then kill off the other 
    streams.&nbsp; This is a tedious process because you have to lookup the 
file 
    list on the stream that failed, compare it to the file list for the streams 
    that are running, kill them, etc.</FONT></P>
    <P><FONT size=2>Does anyone have a better solution?&nbsp; In most cases, 
    logging into the client and using bpbackup is not an option.</FONT> </P>
    <P><FONT size=2>I'm almost tempted to break all of my classes into multiple 
    classes, one stream each, but that just seems ugly (or have other people 
    tried that?).&nbsp; Just rerunning the whole backup is not good because it 
    chews up way too much tape on, say, a 500Gb full ;)&nbsp; </FONT></P>
    <P><FONT size=2>I've been playing with bpbackup on the master server to see 
    if there is some way of saying "backup these files on this client," but 
    bpbackup apparently doesn't support this - it either runs a schedule for a 
    client or you use it on the client.&nbsp; There appears to be no way to say 
    "here's a list of files on client X, now go back them up as if we're 
running 
    a full/incremental/etc."</FONT></P>
    <P><FONT size=2>Anyone have any ideas?</FONT> </P>
    <P><FONT size=2>--</FONT> <BR><FONT size=2>Drew Fabbro </FONT><BR><FONT 
    size=2>fabbro.andrew AT cnf DOT com</FONT> <BR><FONT size=2>503-450-3374 
    (desk)&nbsp; 503-701-0369 (cell)</FONT> <BR><FONT size=2>"Everyone's a 
    superhero.&nbsp; Everyone's a Captain Kirk."</FONT> <BR><FONT 
    size=2>&nbsp;&nbsp; -- Nina Hagen</FONT> 
</P></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C1C4A1.ED3FE190--

<Prev in Thread] Current Thread [Next in Thread>
  • [Veritas-bu] Configurable wait between retries? Was: Thoughts on rerunning a single failed stream?, Fabbro, Andrew P <=