petkovsc
08-02-2005, 22:24
Unless noted, hosts and the server are debian linux 2.4.x. Web browser is
firefox/win32 1.0. All hosts and servers used a clean install of zabbix.
- zabbix_agentd reports to the server it is version 1.1alpha4
- In the web interface, configuration/items, I changed
check_service[http] to check_service[apache], thinking that would get
what I wanted. The status became not supported and the error became not
supported. After reading the manual, I changed it back to
check_service[http]. I enabled it and it continued to work. However
the error remained as not supported, until I cleared it manually from
the database.
- web/config/items. When clicking on a green 'active' status indicator
to change it to not active, then clicking again to change it back to
active, it becomes black, instead of green. Probably because I've
visited the link.
- I'm using syslog-ng which works fine with other programs logging to the
syslog facility. Zabbix_server won't start if I have LogFile commented out. I
don't have any entries in syslog from zabbix. The parent starts long enough
to spawn children, but then everyone dies.
- Diskfree values are not being input into the database as 0. I traced
down the problem to the default configuration. The web interface reports a
custom multiplier of 0. I tried saying do not use, but as soon as I click it,
it takes me back to the item list. However when I look at diskfree, it shows
custom multiplier as 0 again. (I see now, it is because the options change. I
can scroll to the bottom and hit update).
Although the agent is returning KB and the web interface is expecting B.
Changing units to KB gives me "KKB" in my latest data.
As a work around, I changed the multiplier to 1024. This shows me accurate
values and units. This all applies to diskfree, disktotal and diskused. All
had custom multipliers of 0.
- /zabbix-1.1alpha5/include/functions.c: Lines 1099 and 1145
I think this should be update_functions and not update_finctions.
zabbix_log( LOG_LEVEL_DEBUG, "In update_finctions(%d)",item->itemid);
- Default unix host template should change proc_cnt[zabbix_suckerd] and
trapperd to just zabbix_server.
- I have two machines running debian with a 2.6 kernel:
2.6.8-1-686-smp
I'm only using ipv4, however apache2 and postfix are listening on ipv6
ports. I can connect to port 80 and port 25 just fine, however the
net[listen_80] directive does not work. As a work around, I disabled the
net[listen..] and just used check_service. Here is my netstat -nltu:
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 0 0 0.0.0.0:10050 0.0.0.0:* LISTEN
tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN
tcp6 0 0 :::80 :::* LISTEN
tcp6 0 0 :::22 :::* LISTEN
tcp6 0 0 ::1:25 :::* LISTEN
- On reiserfs partitions with the 2.6 kernel inodes returned are 0
(free/total). With a 2.4 kernel it returns 4294967295. I tested one ext3fs
partition and it returned 5998.
- I'm editing a graph with six values incoming and outgoing traffic on eth0
(1min) for three servers. outgoing are bold, incoming are line. Each server
has their own color. Moving a line down works, but few of the ups work.
(only the last one was consistently working). I did have the hosts
interleaved. I wanted them sorted, however most of them wouldn't move. I
worked around it by moving hosts down. Once two hosts showed up next to
eachother, I could move another host down and it would hop over two lines
from the same host. Once I had it in a certain order, the sort order
automatically counted 0,0,1,1,2,2. When they were interleaved most lines
were 0. Some were 1.
firefox/win32 1.0. All hosts and servers used a clean install of zabbix.
- zabbix_agentd reports to the server it is version 1.1alpha4
- In the web interface, configuration/items, I changed
check_service[http] to check_service[apache], thinking that would get
what I wanted. The status became not supported and the error became not
supported. After reading the manual, I changed it back to
check_service[http]. I enabled it and it continued to work. However
the error remained as not supported, until I cleared it manually from
the database.
- web/config/items. When clicking on a green 'active' status indicator
to change it to not active, then clicking again to change it back to
active, it becomes black, instead of green. Probably because I've
visited the link.
- I'm using syslog-ng which works fine with other programs logging to the
syslog facility. Zabbix_server won't start if I have LogFile commented out. I
don't have any entries in syslog from zabbix. The parent starts long enough
to spawn children, but then everyone dies.
- Diskfree values are not being input into the database as 0. I traced
down the problem to the default configuration. The web interface reports a
custom multiplier of 0. I tried saying do not use, but as soon as I click it,
it takes me back to the item list. However when I look at diskfree, it shows
custom multiplier as 0 again. (I see now, it is because the options change. I
can scroll to the bottom and hit update).
Although the agent is returning KB and the web interface is expecting B.
Changing units to KB gives me "KKB" in my latest data.
As a work around, I changed the multiplier to 1024. This shows me accurate
values and units. This all applies to diskfree, disktotal and diskused. All
had custom multipliers of 0.
- /zabbix-1.1alpha5/include/functions.c: Lines 1099 and 1145
I think this should be update_functions and not update_finctions.
zabbix_log( LOG_LEVEL_DEBUG, "In update_finctions(%d)",item->itemid);
- Default unix host template should change proc_cnt[zabbix_suckerd] and
trapperd to just zabbix_server.
- I have two machines running debian with a 2.6 kernel:
2.6.8-1-686-smp
I'm only using ipv4, however apache2 and postfix are listening on ipv6
ports. I can connect to port 80 and port 25 just fine, however the
net[listen_80] directive does not work. As a work around, I disabled the
net[listen..] and just used check_service. Here is my netstat -nltu:
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 0 0 0.0.0.0:10050 0.0.0.0:* LISTEN
tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN
tcp6 0 0 :::80 :::* LISTEN
tcp6 0 0 :::22 :::* LISTEN
tcp6 0 0 ::1:25 :::* LISTEN
- On reiserfs partitions with the 2.6 kernel inodes returned are 0
(free/total). With a 2.4 kernel it returns 4294967295. I tested one ext3fs
partition and it returned 5998.
- I'm editing a graph with six values incoming and outgoing traffic on eth0
(1min) for three servers. outgoing are bold, incoming are line. Each server
has their own color. Moving a line down works, but few of the ups work.
(only the last one was consistently working). I did have the hosts
interleaved. I wanted them sorted, however most of them wouldn't move. I
worked around it by moving hosts down. Once two hosts showed up next to
eachother, I could move another host down and it would hop over two lines
from the same host. Once I had it in a certain order, the sort order
automatically counted 0,0,1,1,2,2. When they were interleaved most lines
were 0. Some were 1.