The 'xx-xxxx.conf' are mine, custom. So I would not have to merge
changes with newer /etc/collectd.conf rpm updates.
I would suggest get a small configuration that is working, set debug
logging[0], and increase the configuration until it fails with little
steps. Load plugin ceph empty, configure then one osd, fails? try mgr?,
try mon? etc
[0] Try something like this?
LoadPlugin logfile
<Plugin logfile>
# not compiled with debug?
# also not writing to the logfile
LogLevel debug
File "/tmp/collectd.log"
# File STDOUT
Timestamp true
# PrintSeverity false
</Plugin>
-----Original Message-----
Cc: [email protected]
Subject: Re: [ceph-users] collectd Ceph metric
Is there any instruction to install the plugin configuration?
Attach my RHEL/collectd configuration file under /etc/ directory.
On RHEL:
[rdma@rdmarhel0 collectd.d]$ pwd
/etc/collectd.d
[rdma@rdmarhel0 collectd.d]$ tree .
.
0 directories, 0 files
[rdma@rdmarhel0 collectd.d]$
I've also checked the collectd configuration file under Ubuntu, there's
no 11-network.conf or 12-memory.conf .etc. However, it could still
collect the cpu and memory information.
On Ubuntu:
nstcc1@nstcloudcc1:collectd$ pwd
/etc/collectd
nstcc1@nstcloudcc1:collectd$ tree .
.
├── collectd.conf
├── collectd.conf.d
│ ├── filters.conf
│ └── thresholds.conf
└── collection.conf
1 directory, 4 files
nstcc1@nstcloudcc1:collectd$
B.R.
Changcheng
On 10:56 Mon 21 Oct, Marc Roos wrote:
>
> Your collectd starts without the ceph plugin ok?
>
> I have also your error " didn't register a configuration callback",
> because I configured debug logging, but did not enable it by loading
> the plugin 'logfile'. Maybe it is the order in which your
> configuration files a read (I think this used to be important with
> collectd)
>
> I have only in my collectd.conf these two lines:
> Include "/etc/collectd.d"
> LoadPlugin syslog
>
> And in /etc/collectd.d/ these files:
> 10-custom.conf (with network section for influx)
> 11-network.conf (ethx)
> 12-memory.conf
> 50-ceph.conf
> 51-ipmi.conf
> 52-ipmi-power.conf
> 53-disk.conf
>
>
>
>
> [0] journalctl -u collectd.service
> Oct 21 10:29:39 c02 collectd[1017750]: Exiting normally.
> Oct 21 10:29:39 c02 systemd[1]: Stopping Collectd statistics daemon...
> Oct 21 10:29:39 c02 collectd[1017750]: collectd: Stopping 5 read
> threads.
> Oct 21 10:29:39 c02 collectd[1017750]: collectd: Stopping 5 write
> threads.
> Oct 21 10:29:40 c02 systemd[1]: Stopped Collectd statistics daemon.
> Oct 21 10:29:53 c02 systemd[1]: Starting Collectd statistics daemon...
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "syslog"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "network"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: Found a configuration for the
> `logfile' plugin, but the plugin isn't loaded or didn't register a
> configuration callback.
> Oct 21 10:29:53 c02 collectd[1031939]: Found a configuration for the
> `logfile' plugin, but the plugin isn't loaded or didn't register a
> configuration callback.
> Oct 21 10:29:53 c02 collectd[1031939]: Found a configuration for the
> `logfile' plugin, but the plugin isn't loaded or didn't register a
> configuration callback.
> Oct 21 10:29:53 c02 collectd[1031939]: network plugin: The
> `MaxPacketSize' must be between 1024 and 65535.
> Oct 21 10:29:53 c02 collectd[1031939]: network plugin: Option
> `CacheFlush' is not allowed here.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "interface"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "memory"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "ceph"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "ipmi"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: ipmi plugin: Legacy
> configuration found! Please update your config file.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "exec"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: plugin_load: plugin "disk"
> successfully loaded.
> Oct 21 10:29:53 c02 collectd[1031939]: Systemd detected, trying to
> signal readyness.
> Oct 21 10:29:53 c02 systemd[1]: Started Collectd statistics daemon.
> Oct 21 10:29:53 c02 collectd[1031939]: Initialization complete,
> entering read-loop.
> Oct 21 10:30:04 c02 collectd[1031939]: ipmi plugin: sensor_list_add:
> Ignore sensor `PS2 Status power_supply (10.2)` of `main`, because it
> is discrete (0x8)! Its type:
> Oct 21 10:30:04 c02 collectd[1031939]: ipmi plugin: sensor_list_add:
> Ignore sensor ` management controller firmware (46.4)` of `main`,
> because it isn't readable! Its Oct 21 10:30:04 c02 collectd[1031939]:
> ipmi plugin: sensor_list_add:
> Ignore sensor ` management controller firmware (46.3)` of `main`,
> because it isn't readable! Its Oct 21 10:30:04 c02 collectd[1031939]:
> ipmi plugin: sensor_list_add:
> Ignore sensor ` management controller firmware (46.2)` of `main`,
> because it isn't readable! Its Oct 21 10:30:04 c02 collectd[1031939]:
> ipmi plugin: sensor_list_add:
> Ignore sensor ` management controller firmware (46.1)` of `main`,
> because it isn't readable! Its Oct 21 10:30:04 c02 collectd[1031939]:
> ipmi plugin: sensor_list_add:
> Ignore sensor `PS1 Status power_supply (10.1)` of `main`, because it
> is discrete (0x8)! Its type:
> Oct 21 10:30:04 c02 collectd[1031939]: ipmi plugin: sensor_list_add:
> Ignore sensor `Chassis Intru system_chassis (23.1)` of `main`, because
> it is discrete (0x5)! Its Oct 21 10:30:04 c02 collectd[1031939]: ipmi
> plugin: sensor_list_add:
> Ignore sensor `HDD Status disk_drive_bay (26.1)` of `main`, because it
> is discrete (0xd)! Its typ Oct 21 10:30:55 c02 collectd[1031939]: ipmi
> plugin: sensor_read_handler:
> sensor `FANB fan_cooling (29.8)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `FANA fan_cooling (29.7)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `FAN6 fan_cooling (29.6)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `FAN5 fan_cooling (29.5)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `FAN1 fan_cooling (29.1)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMMH3 TEMP memory_device (32.94)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMMH2 TEMP memory_device (32.93)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMMG3 TEMP memory_device (32.90)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMMG2 TEMP memory_device (32.89)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMMF3 TEMP memory_device (32.86)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P2-DIMME3 TEMP memory_device (32.82)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P1-DIMMD3 TEMP memory_device (32.78)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P1-DIMMD2 TEMP memory_device (32.77)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P1-DIMMC3 TEMP memory_device (32.74)` of `main` not present.
> Oct 21 10:30:55 c02 collectd[1031939]: ipmi plugin:
sensor_read_handler:
> sensor `P1-DIMMC2 TEMP memory_device (32.73)` of `main` not present.
>
>
>
>
>
> -----Original Message-----
> Cc: [email protected]
> Subject: Re: [ceph-users] collectd Ceph metric
>
> On 10:16 Mon 21 Oct, Marc Roos wrote:
> > I have the same. I do not think ConvertSpecialMetricTypes is
> necessary.
> >
> > <LoadPlugin ceph>
> > Globals true
> > </LoadPlugin>
> >
> > <Plugin ceph>
> > LongRunAvgLatency false
> > ConvertSpecialMetricTypes true
> > <Daemon "osd.1">
> > SocketPath "/var/run/ceph/ceph-osd.1.asok"
> > </Daemon>
> >
> Same configuration, but there's below error after "systemctl restart
> collectd"
> Have you ever hit this error before?
>
> =================Log start===========
> Oct 21 16:22:52 rdmarhel0 collectd[768000]: Found a configuration for
> the `ceph' plugin, but the plugin isn't loaded or didn't register a
> configuration callback.
> Oct 21 16:22:52 rdmarhel0 systemd[1]: Unit collectd.service entered
> failed state.
> Oct 21 16:22:52 rdmarhel0 collectd[768000]: Found a configuration for
> the `ceph' plugin, but the plugin isn't loaded or didn't register a
> configuration callback.
> Oct 21 16:22:52 rdmarhel0 systemd[1]: collectd.service failed.
> Oct 21 16:22:52 rdmarhel0 collectd[768000]: There is a `Daemon' block
> within the configuration for the ceph plugin. The plugin either only
> expects "simple" configuration statements or wasn Oct 21 16:22:52
> rdmarhel0 systemd[1]: collectd.service holdoff time over, scheduling
> restart.
> Oct 21 16:22:52 rdmarhel0 systemd[1]: Stopped Collectd statistics
> daemon.
> -- Subject: Unit collectd.service has finished shutting down
> =================Log end===========
>
> B.R.
> Changcheng
> >
> >
> > -----Original Message-----
> > Cc: [email protected]
> > Subject: Re: [ceph-users] collectd Ceph metric
> >
> > On 09:50 Mon 21 Oct, Marc Roos wrote:
> > >
> > > I am, collectd with luminous, and upgraded to nautilus and
> > > collectd
> > > 5.8.1-1.el7 this weekend. Maybe increase logging or so.
> > > I had to wait a long time before collectd was supporting the
> > > luminous release, maybe it is the same with octopus (=15?)
> > >
> > @Roos: Do you mean that you could run collectd(5.8.1) with
> > Ceph-Nautilus? Below is my collectd configuration with Ceph-Octopus:
> >
> > <LoadPlugin ceph>
> > <Plugin ceph>
> > <Daemon "osd.0">
> > SocketPath "/var/run/ceph/ceph-osd.0.asok"
> > </Daemon>
> > </Plugin>
> >
> > Is there anything wrong?
> >
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: Liu, Changcheng [mailto:[email protected]]
> > > Sent: maandag 21 oktober 2019 9:41
> > > To: [email protected]
> > > Subject: [ceph-users] collectd Ceph metric
> > >
> > > Hi all,
> > > Does anyone succeed to use collectd/ceph plugin to collect ceph
> > > cluster data?
> > > I'm using collectd(5.8.1) and Ceph-15.0.0. collectd failed to
get
> > > cluster data with below error:
> > > "collectd.service holdoff time over, scheduling restart"
> > >
> > > Regards,
> > > Changcheng
> > > _______________________________________________
> > > ceph-users mailing list
> > > [email protected]
> > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
> > >
> > >
> >
> >
>
>
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
