Hello all!
I've posted Release Candidate 1 for Xymon 4.3.20, containing all recent bug fixes since 4.3.19, onto Sourceforge. Reachable at http://sourceforge.net/projects/xymon/files/Xymon/4.3.20/xymon-4.3.20-rc1.ta...
A short list of features/fixes is at https://sourceforge.net/projects/xymon/files/Xymon/4.3.20/, while the full Changelog is at https://sourceforge.net/p/xymon/code/HEAD/tree/branches/4.3.20/Changes
Please take a pounding on it and let me know if you notice any problems -- especially on non-Linux systems. If all goes according to plan, I'll be releasing the final 4.3.20 later on this week.
Regards,
-jc
On 5/12/2015 6:20 AM, J.C. Cleaver wrote:
I've posted Release Candidate 1 for Xymon 4.3.20, . . . Please take a pounding on it and let me know if you notice any problems -- especially on non-Linux systems. If all goes according to plan, I'll be releasing the final 4.3.20 later on this week.
On Solaris 10, xymonnet/c-ares-shim.sh is trying to pass "-e" options to grep and failing during the make.
mv c-ares-1.10.0 c-ares
Must touch "configure", or it will trigger a rebuild because it is older than the tar.gz file.
touch c-ares/configure (cd c-ares && ../c-ares-shim.sh ./configure --disable-shared) grep: illegal option -- e grep: illegal option -- I grep: illegal option -- e grep: illegal option -- D grep: illegal option -- e grep: illegal option -- L Usage: grep -hblcnsviw pattern file . . . ../c-ares-shim.sh: test: argument expected gmake[1]: *** [c-ares/Makefile] Error 1 gmake[1]: Leaving directory `/opt/xymon/xymon-4.3.20-rc1/xymonnet' gmake: *** [xymonnet-build] Error 2
/usr/bin/grep doesn't handle -e /usr/xpg4/bin/grep does
c-ares-shim.sh is just calling 'grep'. Isn't there something in the process which defines $GREP and $EGREP to point to an appropriate version of grep?
Do things because you should, not just because you can.John Thurston 907-465-8591 John.Thurston at alaska.gov Enterprise Technology Services Department of Administration State of Alaska
On Tue, May 12, 2015 11:40 am, John Thurston wrote:
On 5/12/2015 6:20 AM, J.C. Cleaver wrote:
I've posted Release Candidate 1 for Xymon 4.3.20, . . . Please take a pounding on it and let me know if you notice any problems
especially on non-Linux systems. If all goes according to plan, I'll be releasing the final 4.3.20 later on this week.
On Solaris 10, xymonnet/c-ares-shim.sh is trying to pass "-e" options to grep and failing during the make.
mv c-ares-1.10.0 c-ares
Must touch "configure", or it will trigger a rebuild because it is
older than the tar.gz file. touch c-ares/configure (cd c-ares && ../c-ares-shim.sh ./configure --disable-shared) grep: illegal option -- e grep: illegal option -- I grep: illegal option -- e grep: illegal option -- D grep: illegal option -- e grep: illegal option -- L Usage: grep -hblcnsviw pattern file . . . ../c-ares-shim.sh: test: argument expected gmake[1]: *** [c-ares/Makefile] Error 1 gmake[1]: Leaving directory `/opt/xymon/xymon-4.3.20-rc1/xymonnet' gmake: *** [xymonnet-build] Error 2
/usr/bin/grep doesn't handle -e /usr/xpg4/bin/grep does
c-ares-shim.sh is just calling 'grep'. Isn't there something in the process which defines $GREP and $EGREP to point to an appropriate version of grep?
Unfortunately not -- configure hadn't had a need for it it seems, and we're not within autoconf currently.
Can you try this quick patch to it? Although uglier, I don't think there's a need for a single (e)grep there.
-jc
On 5/12/2015 6:20 AM, J.C. Cleaver wrote: ...
Please take a pounding on it and let me know if you notice any problems --
combostatus doesn't seem to be working as expected.
Lines in combo.cfg of the form: f.bar.com.conn = foo.bar.com.conn work just fine.
Lines of the form: f.bar.com.conn = foo.bar.com.conn + baz.bar.com.conn also seem to work fine.
But lines containing any constants seem to fail: f.bar.com.conn = foo.bar.co.connm + 1 leave footprints in the debug log like:
2015-05-12 13:37:38.838687 Item '1' has no testname part 2015-05-12 13:37:38.838886 Item '1' has no testname part 2015-05-12 13:37:38.839026 Invalid data for symbol calculation - missing host/testname: 1 26080 2015-05-12 13:37:38.839158 Symbolic 'foo.bar.com.conn+1' converted to '1+0'
and things like: f.bar.com.conn = (foo.bar.com.conn + baz.bar.com.conn ) >= 1 leave footprints in the debug log like:
2015-05-12 13:37:38.839331 Item '1' has no testname part 2015-05-12 13:37:38.839476 Item '1' has no testname part 2015-05-12 13:37:38.839609 Invalid data for symbol calculation - missing host/testname: 1 26080 2015-05-12 13:37:38.839775 Symbolic '(foo.bar.com.conn+baz.bar.com.conn)>=1' converted to '(1+1)>=0'
It seems to assume everything must be a host.test name-pair doesn't seem able to parse the constants off the line. Yes, this is still under Solaris 10 on sparc.
-- Do things because you should, not just because you can.
John Thurston 907-465-8591 John.Thurston at alaska.gov Enterprise Technology Services Department of Administration State of Alaska
On Tue, May 12, 2015 2:44 pm, John Thurston wrote:
On 5/12/2015 6:20 AM, J.C. Cleaver wrote: ...
Please take a pounding on it and let me know if you notice any problems
combostatus doesn't seem to be working as expected.
Lines in combo.cfg of the form: f.bar.com.conn = foo.bar.com.conn work just fine.
Lines of the form: f.bar.com.conn = foo.bar.com.conn + baz.bar.com.conn also seem to work fine.
But lines containing any constants seem to fail: f.bar.com.conn = foo.bar.co.connm + 1 leave footprints in the debug log like:
2015-05-12 13:37:38.838687 Item '1' has no testname part 2015-05-12 13:37:38.838886 Item '1' has no testname part 2015-05-12 13:37:38.839026 Invalid data for symbol calculation - missing host/testname: 1 26080 2015-05-12 13:37:38.839158 Symbolic 'foo.bar.com.conn+1' converted to '1+0'
and things like: f.bar.com.conn = (foo.bar.com.conn + baz.bar.com.conn ) >= 1 leave footprints in the debug log like:
2015-05-12 13:37:38.839331 Item '1' has no testname part 2015-05-12 13:37:38.839476 Item '1' has no testname part 2015-05-12 13:37:38.839609 Invalid data for symbol calculation - missing host/testname: 1 26080 2015-05-12 13:37:38.839775 Symbolic '(foo.bar.com.conn+baz.bar.com.conn)>=1' converted to '(1+1)>=0'
It seems to assume everything must be a host.test name-pair doesn't seem able to parse the constants off the line. Yes, this is still under Solaris 10 on sparc.
Hi,
I suspect this is actually related to https://sourceforge.net/p/xymon/code/7610/ (adding initial-digit support for hostnames to combo.cfg). Taking a closer look now.
Regards,
-jc
On 5/12/2015 6:20 AM, J.C. Cleaver wrote: ...
Please take a pounding on it and let me know if you notice any problems --
Possibly related to my earlier reported combostatus errors, my web display now includes an extraneous column for the line for my xymon server. In addition to the expected columns (cpu, disk, files, etc), there is now one titled "/opt/xymon/server/bin/combostatus" The color is red.
-- Do things because you should, not just because you can.
John Thurston 907-465-8591 John.Thurston at alaska.gov Enterprise Technology Services Department of Administration State of Alaska
On 5/12/2015 2:03 PM, John Thurston wrote:
On 5/12/2015 6:20 AM, J.C. Cleaver wrote: ...
Please take a pounding on it and let me know if you notice any problems --
Possibly related to my earlier reported combostatus errors, my web display now includes an extraneous column for the line for my xymon server. In addition to the expected columns (cpu, disk, files, etc), there is now one titled "/opt/xymon/server/bin/combostatus" The color is red.
This is definitely related. After editing my combostatus config file to remove all lines with constant numbers, the column has gone purple. Fixing the underlying parsing problem will probably make this one go away, too.
-- Do things because you should, not just because you can.
John Thurston 907-465-8591 John.Thurston at alaska.gov Enterprise Technology Services Department of Administration State of Alaska
participants (2)
-
cleaver＠terabithia.org
-
john.thurston＠alaska.gov