Yann,
I'm suspecting someting wrong with the way robinhood indexes entries for
this particular filesystem type (cifs).
Robinhood daily testing with POSIX filesystems is done on ext3/4 and
xfs, and most sites using the POSIX mode run it on nfs, fhgfs, ...
So I think we have to look into this direction.
On POSIX fs, robinhood mainly uses the inode number to address entries.
To get the inode number of '/mnt/bio/GBarazzone' you can run:
"ls -lid /mnt/bio/GBarazzone"
> can you run rbh-du in FULL mode, so we can see what 'id' it tries to
use for addressing 'GBarazzone'
rbh-du -l FULL -dH /mnt/bio/GBarazzone
> Run a partial scan on "/mnt/bio/GBarazzone" in FULL debug mode. We
will see what 'id' it inserts into the database.
robinhood -f bioimagerie --scan=/mnt/bio/GBarazzone --once --no-gc
Thanks,
Thomas
On 02/16/16 17:22, Yann Sagon wrote:
5)
rbh-du -l VERB -dH /mnt/bio/GBarazzone
Using config file '/etc/robinhood.d/tmpfs/bioimagerie.conf'.
2016/02/16 17:17:02 [3193/1] CheckFS | '/mnt/bio' matches mount point
'/mnt/bio', type=cifs, fs=\\\\10.40.44.30\\bioimagerie
2016/02/16 17:17:02 [3193/1] ListMgr | Accounting stats not available
2016/02/16 17:17:02 [3193/1] du | Notice: no attrs in DB for
/mnt/bio/GBarazzone
I thought it was ok because I've only tested rbh-du using point 2
which seems correct.
Point 3 is incorrect. Point 4 count the files in /mnt/bio/GBarazzone
and the result is clearly different. In point 3, there are no line
showing the file count.
Does point 5 gives a hint?
Should I try rbh 3 and forget this version?
Thanks
2016-02-16 15:21 GMT+01:00 Yann Sagon <[email protected]
<mailto:[email protected]>>:
Hello,
I'm not using %include.
Here are my two ListManager block:
ListManager
{
commit_behavior = transaction;
user_acct = off;
group_acct = off;
MySQL
{
#server = localhost;
#port = 3306;
db = rbh_scinas_biomagerie;
user = robinhood;
password_file = /etc/robinhood.d/.dbpassword;
socket = "/var/lib/mysql/mysql.sock" ;
}
}
ListManager
{
commit_behavior = transaction;
user_acct = off;
group_acct = off;
MySQL
{
#server = localhost;
#port = 3306;
db = rbh_baobab_home;
user = robinhood;
password_file = /etc/robinhood.d/.dbpassword;
socket = "/var/lib/mysql/mysql.sock" ;
}
}
2016-02-16 15:13 GMT+01:00 LEIBOVICI Thomas
<[email protected] <mailto:[email protected]>>:
Hi Yann,
There shouldn't be any mix between the 2 robinhood instances
as long as you have 2 distinct databases.
Isn't there any issue that could be hidden by an "%include",
or caused by a duplicated ListManager config block?
Thomas
On 02/16/16 14:52, Yann Sagon wrote:
update:
I have stopped robinhood, removed the second conf file,
restarted robinhood and now the behaviour is correct.
I suspect something happen with the db. I'm using same
user/pass for both conf but distinct db. Is this a problem or
is this a bug?
Best
2016-02-15 17:05 GMT+01:00 Yann Sagon <[email protected]
<mailto:[email protected]>>:
Hello, any hint? I have dropped the db, recreated it
(using rbh-config) and relaunched robinhood.
The result is the same. Is there any other place where I
can clean something?
For clarification: I have two conf in
/etc/robinhood.d/tmpfs and when I start robinhood, both
of them are run. The problematic one terminate before the
correct one.
Is this possible that there is a mix between them? They
use of course a separate db (but with same user/pass).
Thanks
2016-02-11 16:14 GMT+01:00 Yann Sagon
<[email protected] <mailto:[email protected]>>:
Hello Thomas,
I see nothing special in the log. I have extracted
the logs of today run.
Thanks for any suggestion
2016-02-10 20:42 GMT+01:00 Thomas LEIBOVICI (mail
perso) <[email protected]
<mailto:[email protected]>>:
Hi Yann,
Perhaps something wrong occured at the end of the
scan - like a wrong garbage collection of DB entries.
Did you find something in the robinhood log, to
see if something wrong occured during the scan?
Regards,
Thomas
On 10/02/2016 17:10, Yann Sagon wrote:
Hello,
I'm using version 2.5.5 of robinhood on a cifs
filesystem.
I'm using tmpfs. I have two conf files
beegfs.conf
and
bioimagerie.conf
When I launch robinhood, I can do something like
that:
rbh-du -f bioimagerie -Hd /mnt/bio/_BACKUPS
Using config file
'/etc/robinhood.d/tmpfs/bioimagerie.conf'.
/mnt/bio/_BACKUPS
dir count:296108, size:0, spc_used:0
file count:5031540, size:9.3T, spc_used:12.8T
(the numbers aren't correct, I don't have an ok
output at hand).
When the scan is completed, the same command
shows this:
rbh-du -f bioimagerie -Hd /mnt/bio/_BACKUPS
Using config file
'/etc/robinhood.d/tmpfs/bioimagerie.conf'.
/mnt/bio/_BACKUPS
dir count:1, size:0, spc_used:0
Do you have any clue? I *think* it was working
when I had only one conf file. It's working for
the other conf file. The two conf files are
almost identical. Only the db, log and path are
changing.
Should I try to do a rbh-config empty_db ?
Thanks
--
Yann SAGON
Ingénieur système HPC
24 Rue du Général-Dufour
1211 Genève 4 - Suisse
Tél. : +41 (0)22 379 7737
<tel:%2B41%20%280%2922%20379%207737>
[email protected] <mailto:[email protected]> -
www.unige.ch <http://www.unige.ch>
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
_______________________________________________
robinhood-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/robinhood-support
