11 Apr
2012
11 Apr
'12
3:16 p.m.
These are synced from an array at the end if the poller function.
The array is built from each module as it updates the rrd. So this is a particularly weird one!
Adam.
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.
Chris Stone axisml@gmail.com wrote:
Had some time to look at this more indepth and with /debug=yes, took
the query just after the Graphs label and ran it manually against the
database and found:
mysql> SELECT * FROM device_graphs WHERE device_id = '65'
-> ;
Empty set (0.00 sec)
OOPS - there were some there at one point. Checked a similar server and found:
mysql> SELECT * FROM device_graphs WHERE device_id = '55';
+-----------+-------------------+
| device_id | graph |
+-----------+-------------------+
| 55 | uptime |
| 55 | netstat_icmp |
| 55 | netstat_icmp_info |
| 55 | netstat_snmp |
| 55 | netstat_snmp_pkt |
| 55 | hr_processes |
| 55 | hr_users |
| 55 | ucd_cpu |
| 55 | ucd_swap_io |
| 55 | ucd_io |
| 55 | ucd_contexts |
| 55 | ucd_interrupts |
| 55 | ucd_memory |
| 55 | ucd_load |
| 55 | netstat_ip |
| 55 | netstat_ip_frag |
| 55 | netstat_tcp |
| 55 | netstat_udp |
+-----------+-------------------+
18 rows in set (0.00 sec)
So, inserted the uptime one back in for the server where the graphs
are missing (id 65) and refreshed the page - uptime graph is back with
no gaps in the graphs. Inserted the records for the others and they
are now all back. Had another server with the same problem - it's
fixed now too.
So, not sure what did it, but seems that at some point, apparently
with an update since no other access of the database directly has been
done, these records were deleted from the device_graphs table.
Chris
_____________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium