Amanda-Users

Re: Amanda-2.4.3b4 and tar exclude lists working?

2002-09-03 20:09:04
Subject: Re: Amanda-2.4.3b4 and tar exclude lists working?
From: Jean-Francois Malouin <Jean-Francois.Malouin AT bic.mni.mcgill DOT ca>
To: Frank Smith <fsmith AT hoovers DOT com>
Date: Tue, 3 Sep 2002 15:56:34 -0400
* Frank Smith (fsmith AT hoovers DOT com) [20020903 12:53] thus spake:

[...]

> > ---
> > exclude file
> > ---
> >
> > [root@blade]# cat .amanda-gnutar-exclude-list.txt
> > ./diffusion/
> > ./extra-TLE/
> > ./F_MRI/
> > ./MRI_misc/
> > ./MT/
> > ./neda/
> > ./PREVIOUS/
> > ./relaxo/
> > ./restore/
> > ./samson/
> > ./thalamus/
> > ./TLE/
> >
> >
> 
> Yes, it looks like you have it set up correctly, maybe it is a 2.4.3b4
> thing (I'm still running 2.4.2p2).  What issues are you seeing with
> the latest version?

some chunk is way beyond tape capacity, but with the exclude list it
should fit nicely. Looking at my logs I see that somehow, when amanda
makes the schedule it does not take into account the exclude list,
miscalculate the size of the chunk and so complains that it can't be
backed up because it's too big:

---
amandad debug file contains:
---

[...]

Amanda 2.4 REQ HANDLE 000-18320808 SEQ 1031029801
SECURITY USER amanda
SERVICE sendsize
OPTIONS features=fffffeff9f00;maxdumps=1;hostname=blade;

[...]

GNUTAR /raid/epilepsy 0 1970:1:1:0:0:0 -1 OPTIONS 
|;auth=bsd;index;exclude-list=.amanda-gnutar-exclude-list.txt; 

[...]

Amanda 2.4 REP HANDLE 000-18320808 SEQ 1031029801
OPTIONS features=fffffeff9f00;
/raid/epilepsy 0 SIZE 35162160

---

sendsize debug file contains:

sendsize: debug 1 pid 24383 ruid 33 euid 33: start at Tue Sep  3
01:10:00 2002
sendsize: version 2.4.3b4-20020829
sendsize[24383]: time 0.480: waiting for any estimate child
sendsize[24510]: time 0.481: calculating for amname '/raid/epilepsy', dirname 
'/raid/epilepsy', spindle -1
sendsize[24510]: time 0.481: getting size via gnutar for /raid/epilepsy level 0
sendsize[24510]: time 0.492: spawning /opt/amanda_vxa/libexec/runtar in pipeline
sendsize[24510]: argument list: /bin/gtar --create --file /dev/null --directory 
/raid/epilepsy --one-file-system --listed-incremental 
/opt/amanda_vxa/var/amanda/gnutar-lists/blade_raid_epilepsy_0.new --sparse 
--ignore-failed-read --totals --exclude-from 
/tmp/amanda/sendsize._raid_epilepsy.20020903011000.exclude .
sendsize[24510]: time 7.731: Total bytes written: 36006051840 (33GB, 4.2GB/s)
sendsize[24510]: time 7.732: .....
sendsize[24510]: estimate time for /raid/epilepsy level 0: 7.240
sendsize[24510]: estimate size for /raid/epilepsy level 0: 35162160 KB
sendsize[24510]: time 7.732: waiting for /bin/gtar "/raid/epilepsy"
child
sendsize[24510]: time 7.732: after /bin/gtar "/raid/epilepsy" wait
sendsize[24510]: time 7.733: done with amname '/raid/epilepsy', dirname 
'/raid/epilepsy', spindle -1
sendsize[24383]: time 7.733: child 24510 terminated normally

---
Notice above the line "Total bytes written: 36006051840 (33GB..."
---

---
The file from the above output --exclude-from" contains my exclude file list:
---
[root@blade]# cat /tmp/amanda/sendsize._raid_epilepsy.20020903011000.exclude

./diffusion/
./extra-TLE/
./F_MRI/
./MRI_misc/
./MT/
./neda/
./PREVIOUS/
./relaxo/
./restore/
./samson/
./thalamus/
./TLE/

My exclude list!
---

---
runtar output
---

[root@blade]# cat /tmp/amanda/runtar.20020903011000.debug
runtar: debug 1 pid 24519 ruid 33 euid 0: start at Tue Sep  3 01:10:00 2002 
/bin/gtar: version 2.4.3b4-20020829 running: /bin/gtar: 
/bin/gtar --create --file /dev/null --directory /raid/epilepsy 
--one-file-system --listed-incremental 
/opt/amanda_vxa/var/amanda/gnutar-lists/blade_raid_epilepsy_0.new --sparse 
--ignore-failed-read --totals --exclude-from 
/tmp/amanda/sendsize._raid_epilepsy.20020903011000.exclude . 


> Unrelated question:
> Doesn't the trailing slash on your exclude list entries mean that only
> those directories themselves are skipped and not any subdirectories?
> I leave off the slash to prune the directory tree at that point, and
> can't think of too many uses for excluding a directory while backing
> up its subdirectories.  Or possibly my idea of exclude syntax is off.

I sure don't know! But I suspect that you're right.
Actually I shoud rather have something like

./blah/*

Thanks for the time,
jf

> 
> Frank
> 
-- 
Thought to justify wrong doings, and speech to conceal thoughts.

<Prev in Thread] Current Thread [Next in Thread>