Reversing the effects of a 'bb 127.0.0.1 "drop HOSTNAME TEST"' command
On Wed, Mar 2, 2011 at 4:21 PM, Ralph Mitchell <ralphmitchell at gmail.com>wrote:
On Wed, Mar 2, 2011 at 7:37 PM, Jonathan B. Horen <jbhoren at alaska.edu>wrote:
!@#$%
I've never been good a dealing with the aftermath of purple alerts, and always end-up using brute-force to remove all hostname references in $XYMON/data. But this time I really screwed-up and ran
bb 127.0.0.1 "drop HOSTNAME TEST"
for the purple-affected hosts... and now I can't get 'em back!
I could be wrong, but I think you're pretty much done, as far as the old information goes. Actually, I'm not sure about the RRDs. However, whatever was purple will come back just as soon as the next report is delivered. If you want to put the dots back up manually, try this:
bb 127.0.0.1 "status hostname,domain,com.stuff green"to send a green status for the "stuff" column. Those are commas in the hostname, because back in the day BB used the period to separate the test name from the host name. Give it 30 mins and it'll go purple all on its own... :)
You could follow up with:
bb 127.0.0.1 "disable hostname,domain,com.stuff -1 wups, broke it"to send "disable until OK", so those pesky purple dots stay away.
I guess I didn't express myself correctly. I need the tests to restart, not to have a green status-LED place-holder.
At this point, I've:
- stopped xymon on the server
- stopped xymon on the three affected clients
- copied all of $XYMON/data/{hist,hostdata,hostlogs,rrd} from my last backup to $XYMON/data
- started xymon on the server
- started xymon on the three affected clients
No joy. The LEDs for cpu, disk, files, memory, msgs, ports, and procs on the three affected hosts remain missing.
On each of the affected clients, $XYMON/tmp/{hobbit_vmstat.biotech.nnn,msg.biotech.txt) exist and are updated regularly (@5-minute intervals).
So, what's wrong on the server? Why isn't the information being generated-and-displayed on bb.html?
What am I missing?
-- JONATHAN B. HOREN ARSC/LSI Systems Administrator WRRB/008-001 T: (907) 450-8694 E: jbhoren at alaska.edu *"After Tuesday, even the calendar says W T F!!"*
Den 03-03-2011 18:46, Jonathan B. Horen skrev:
I guess I didn't express myself correctly. I need the tests to restart, not to have a green status-LED place-holder.
At this point, I've:
* stopped xymon on the server * stopped xymon on the three affected clients * copied all of $XYMON/data/{hist,hostdata,hostlogs,rrd} from my last backup to $XYMON/data * started xymon on the server * started xymon on the three affected clientsNo joy. The LEDs for cpu, disk, files, memory, msgs, ports, and procs on the three affected hosts remain missing.
Your clients are not sending data to the Xymon server. If they did, then the cpu, disk etc. statuses would show up right away.
On each of the affected clients, $XYMON/tmp/{hobbit_vmstat.biotech.nnn,msg.biotech.txt) exist and are updated regularly (@5-minute intervals).
So, what's wrong on the server? Why isn't the information being generated-and-displayed on bb.html?
Check the "Reports->Ghost clients" display to see if the hosts are reporting with another name than what's in the hosts.cfg / bb-hosts file.
And check the logfiles on the clients to see if there are problems sending data over to the Xymon server.
Regards, Henrik
participants (2)
-
henrik@hswn.dk
-
jbhoren@alaska.edu