Hello,
This is Jane Liu from BNL. We experienced a Robinhood client issue
recently and could not get through. I am writing to seek for help. Here
are details.
We have a dedicated MDS(+MGS) server named sphnxmds01 (RHEL8) in our
Lustre. We deployed a dedicated Robinhood server (RHEL8) named
sphnxrbh01. It works well and I could run the command as below.
[root@sphnxrbh01 ~]# rbh-report -i
Using config file '/etc/robinhood.d/myfs.conf'.
type, count, volume, spc_used, avg_size
symlink, 17715, 1.73 MB, 69.18 MB, 102
dir, 388567, 2.76 GB, 2.77 GB, 7.46 KB
file, 19872060, 42.74 PB, 35.69 PB, 2.26 GB
Total: 20278342 entries, volume: 48122437338606661 bytes (42.74 PB),
space used: 40180446982512640 bytes (35.69 PB)
Now we plan to allow users to query Robinhood service from a client
host. So I deployed Robinhood client tool in CVMFS and tried to run
rbh-report on a client host phnxsub01 (RHEL7) . However it showed
nothing when running rbh-report. Please see the output, config and logs
below:
[root@phnxsub01 /]#
/cvmfs/sdcc.bnl.gov/software/robinhood/sbin/rbh-report -i -f
/etc/robinhood.d/myfs.conf
[root@phnxsub01 /]# cat /etc/robinhood.d/myfs.conf
mode: c; c-basic-offset: 4; indent-tabs-mode: nil;
vim:expandtab:shiftwidth=4:tabstop=4:
General
{ fs_path = "/mnt"; # filesystem type, as displayed by 'mount'
(e.g. ext4, xfs, lustre, ...) fs_type = lustre; }
Log { # Log verbosity level # Possible values are: CRIT, MAJOR,
EVENT, VERB, DEBUG, FULL debug_level = EVENT; # Log file
log_file = "/var/log/lustre.log"; # File for reporting purge
events report_file = "/var/log/lustre_actions.log"; # set
alert_file, alert_mail or both depending on the alert method you wish
alert_file = "/var/log/lustre_alerts.log"; alert_show_attrs = yes;
}
ListManager {
MySQL { server = sphnxrbh01.sdcc.bnl.gov; db =
robinhood_sphnx01; user = robinhood; password_file =
/etc/robinhood.d/.dbpassword; }
}
# Lustre 2.x only
ChangeLog {
MDT { mdt_name = "MDT0000"; reader_id = "cl2"; }
}
#%include "includes/modeguard.inc"
%include "/etc/robinhood.d/includes/modeguard.inc"
[root@phnxsub01 /]# cat /etc/robinhood.d/includes/modeguard.inc
# Robinhood Policy definition for modeguard
define_policy modeguard { status_manager = modeguard; scope \{
type == directory }
default_action = modeguard.enforce_mode;
default_lru_sort_attr = last_mod;
}
[root@phnxsub01 /]# tail /var/log/lustre.log
2023/10/16 16:33:08 [42486/1] CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'
2023/10/16 16:33:08 [42486/1] CheckFS | Drop the database and restart
the daemon.
2023/11/07 12:18:52 [24388/1] CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'
2023/11/07 12:18:52 [24388/1] CheckFS | Drop the database and restart
the daemon.
2023/11/07 12:25:05 [24812/1] CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'
2023/11/07 12:25:06 [24812/1] CheckFS | Drop the database and restart
the daemon.
2023/11/07 14:35:00 [33709/1] CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'
2023/11/07 14:35:00 [33709/1] CheckFS | Drop the database and restart
the daemon.
2023/11/07 14:51:47 [34559/1] CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'
2023/11/07 14:51:47 [34559/1] CheckFS | Drop the database and restart
the daemon.
The error says "CheckFS | Root id changed! Previous value:
'0x200000007:0x1:0x0', new value: '57934691/144115188193296385'". What
does this mean? I don't want to drop the database since robinhood client
commands work perfectly on sphnxrbh01. Could you help to check what have
caused the error message? thanks.
For debugging, here is myfs.conf on Robinhood server sphnxrbh01.
[root@sphnxrbh01 ~]# cat /etc/robinhood.d/myfs.conf
# mode: c; c-basic-offset: 4; indent-tabs-mode: nil;
# vim:expandtab:shiftwidth=4:tabstop=4:
General { fs_path = "/mnt"; # filesystem type, as displayed by
'mount' (e.g. ext4, xfs, lustre, ...) fs_type = lustre; }
Log
{ # Log verbosity level # Possible values are: CRIT, MAJOR,
EVENT, VERB, DEBUG, FULL debug_level = EVENT; # Log file
log_file = "/var/log/lustre.log"; # File for reporting purge
events report_file = "/var/log/lustre_actions.log"; # set
alert_file, alert_mail or both depending on the alert method you wish
alert_file = "/var/log/lustre_alerts.log"; alert_show_attrs = yes;
}
ListManager {
MySQL
{ server = sphnxrbh01.sdcc.bnl.gov; db =
robinhood_sphnx01; user = robinhood; password_file =
/etc/robinhood.d/.dbpassword; }
}
Lustre 2.x only
ChangeLog {
MDT
{ mdt_name = "MDT0000"; reader_id = "cl2"; }
}
%include "includes/modeguard.inc"
Thanks,
Jane
_______________________________________________
robinhood-support mailing list
robinhood-support@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/robinhood-support
