But question N2 is still actual: How should I (or should I?) merge data for under old and new hostnames under data/hostdata ? will it be just enough move content of data/hostdata/oldhost/* to data/hostdata/newhost/* ?
By the way, I had found on both XyMon servers there are still directories for old nad new hostnames under xymon/data/hostdata/ old host data not updated any more, but new ones only. Question N2: Is it expected or after xymon "rename ... " call all should be merged into one directory per host?
Andrey Chervonets/Cominder/LV wrote on 2013.07.11 09:30:54:
From: Andrey Chervonets/Cominder/LV To: xymon at xymon.com, Date: 2013.07.11 09:30 Subject: Re: Strange behaviour: summary reports purple while test details is OK
At last I had found solution myself:
stop XyMon both on client and server and remove all files under XyMon tmp directory: alert.chk alert.chk.sub xymond.chk and others
I am not sure which exactly file was wrong ( I suppose it was xymond.chk on server), but now it is working. The only backside effect - all metrics status last change time reset so all metric images are with "round" smiling gifs.
Best regards,
Andrey Chervonets
SIA CoMinder http://www.cominder.eu/
Andrey Chervonets/Cominder/LV wrote on 2013.07.08 19:32:54:
From: Andrey Chervonets/Cominder/LV To: xymon at xymon.com, Date: 2013.07.08 19:32 Subject: Strange behaviour: summary reports purple while test details is OK
Good day!
We had renamed 2 hosts today and renamed all (I hope) at host level and in XyMon config and restarted XyMon server and clients too. Rename in XyMon was done as: a) changes records in hosts.cfg b) xymon 127.0.0.1:port "rename OLD_HOST NEW_HOST"
Now XyMon Server (4.3.11) - generates purple for 1st host in non- green and group reports while every test details shows green, yellow, red and so on what is expected. the same server - generates normal reports for 2nd host renames
We have 2 XyMon server for High Availability and clients send info to both. The 2nd XyMon Server (4.3.4) - generate normal reports for both renamed hosts.
So, question N 1 is - how to find where is the problem? This only thing I had found in info report (which with trends are always green by the way) is that different status for the same metric is reported with minimal interval delta (~ 10 min which may be time since hostname change and xymon client restart and when "rename" was called at XyMon server. For example: Status summary:
Color Service Since Duration
greed cpu Mon Jul 08 14:09:28 2013 0 days, 05 hours, 02 minutes purple cpu Mon Jul 08 13:59:44 2013 0 days, 05 hours, 12 minutes greed disk Mon Jul 08 18:24:07 2013 0 days, 00 hours, 48 minutes purple disk Mon Jul 08 13:59:44 2013 0 days, 05 hours, 12 minutes yellow files Mon Jul 08 14:09:28 2013 0 days, 05 hours, 02 minutes purple files Mon Jul 08 13:59:44 2013 0 days, 05 hours, 12 minutes greed memory Mon Jul 08 14:09:28 2013 0 days, 05 hours, 02 minutes purple memory Mon Jul 08 13:59:44 2013 0 days, 05 hours, 12 minutes greed msgs Mon Jul 08 14:14:29 2013 0 days, 04 hours, 57 minutes purple msgs Mon Jul 08 13:59:44 2013 0 days, 05 hours, 12 minutes
This mean XyMon counts summary for 2 hostnames - new one (with info received just after hostname rename) and old one, but renamed to new name.