PROCS not monitored correctly -what is wrong with this picture?? (Find the error!) (LONG) 🔗 link

My apologies to anyone who gets this on a digest
I'm having some odd issues with procs on a database cluster and I would be
very grateful for any clues!


We have multiple rule sets to cut down on multiple alerts for shared disk,
and because some have more capacity

db0-db6 are Solaris 10 servers (there is no db8) running xymon 4.3.10
db11-db62 are RHEL Linux running a mix of xymon 4.3.10 and 4.3.7
server is a RHEL VM running xymon 4.3.10

Procs have been screwey for a while, but we only noticed when we added rngd
testing to the linux boxes

db1, db2, db3, db6 are alerting for missing rngd . db0, db4 and db5 are NOT
(this is particularly puzzling since db5&db6 share a rule set, as do db0
and db1. I'd expect them to go as pairs.

db1, db2, db3, and db6 all show:

[image: green] cron (found 1, req. 1 or more)[image: green] nscd
(found 1, req. 1 or more)[image: green] xntpd (found 1, req. 1 or
more)[image: green] cron (found 1, req. 1 or more)[image: green] nscd
(found 1, req. 1 or more)[image: green] ntpd (found 1, req. 1 or
more)[image: yellow] rngd (found 0, req. 1 or more)[image: green] cron
(found 1, req. 1 or more)


db0, db4 and db5 all show:

[image: green] cron (found 1, req. 1 or more)[image: green] nscd
(found 1, req. 1 or more)[image: green] xntpd (found 1, req. 1 or
more)[image: green] cron (found 1, req. 1 or more)


That last cron comes from a HOST=* rule atthe  end.

Can you find what is wrong with the lines below?
#-------------------#Database Servers
HOST=%^db[0|1].example.com
        MEMPHYS 100 101
        MEMSWAP 85 95
        PROC cron 1 -1 yellow
        PROC nscd 1 -1 yellow
        PROC xntpd 1 -1 yellow
#        PROC sar -1 4 yellow
        LOAD 80.0 120.0
        DISK /oracle/dba_msc_nfs2 101 101
        DISK /oracle/data09 99 99
        DISK /oracle/data10 97 98
        DISK /oracle/data17 96 97
        DISK /oracle/data22 98 99
        DISK %.*archivelogs.* 90 95
        DISK %.*redologs.* 90 95
        DISK %.*data.* 95 96
        LOG /export/home/xymon/client/tmp/powermt.out %degraded COLOR=yellow

HOST=%^db[2|3|4|8].example.com
        UP 30m 9999d
        MEMPHYS 100 101
        MEMSWAP 85 95
#       LOAD    48.0 64.0
        LOAD 80.0 120.0
        PROC cron 1 -1 yellow
        PROC nscd 1 -1 yellow
        PROC xntpd 1 -1 yellow
#        PROC sar -1 4 yellow
        DISK %.*data.* IGNORE
        DISK %.*oracle.* IGNORE
        DISK %.*redologs.* IGNORE
        DISK %.* 80 90
        LOG /export/home/xymon/client/tmp/powermt.out %degraded COLOR=yellow


HOST=%^db[5|6].example.com
        UP 30m 9999d
        MEMPHYS 100 101
        MEMSWAP 85 95
        LOAD    160.0 240.0
        DISK /oracle/export02 90 95
        DISK %.*oracle.* IGNORE
        DISK %.* 80 90
        PROC cron 1 -1 yellow
        PROC nscd 1 -1 yellow
        PROC xntpd 1 -1 yellow
#        PROC sar -1 4 yellow
        LOG /export/home/xymon/client/tmp/powermt.out %degraded COLOR=yellow

HOST=%^db[11|12|13|21|22|23|31|32|33|61|62].bo3.*
        UP 30m 9999d
        MEMPHYS 100 101
        MEMSWAP 85 95
        LOAD    64.0 128.0
        PROC cron 1 -1 yellow
        PROC nscd 1 -1 yellow
        PROC ntpd 1 -1 yellow
        PROC rngd 1 -1 yellow


(there are usually more LOG lines but I removed them for clarity, and
verified that the errors persist)

heh and as soon as I sent that whole LONG thing, someone on my team points
out a clue which leads to
HOST=%^db[11|12|13|21|22|23|31|32|33|61|62].bo3.*

which is of course not a well-formed regexp. I dont' see why it would match
db1.bo3.example.com, but it's certainly mucked up.

Yes all of your "or" expressions should be in parenthesis, not brackets,
what you're saying with the brackets is "match any character contained
within the brackets", so the '|' isn't being interpreted as an "or" but
just as a pipe character to match. So the regular expression ^db[11|12|12
...] is saying "match any phrase that starts with db and has 1 or 1 or | or
1 or 2 ... as a second character, so db1.bo3.example.com certainly matches.

Thanks,
Larry Barber

On Tue, Apr 9, 2013 at 1:28 PM, Betsy Schwartz <user-c61747246f66@xymon.invalid>wrote:

Yep, thanks again. I fixed the regexp,  but I also hadn't realized that "|"
is not a special character within brackets, so I fixed it to the
not-really-correct but working:

HOST=%^db[1|2|3|6][1|2|3].bo3.*

which works but has extraneous characters, it should be:

HOST=%^db[1236][123].bo3.*

or to get exactly what I had before:

HOST=%^db(11|12|13|21|22|23|31|32|33|61|62).bo3.*

Although for our particular purposes I could also just do:

HOST=%^db[1-6][0-9].bo3.*

since that matches all the present and future dbN.bo3 hosts except db0-db9
, which are special

educational day!

HOST=%^db[11|12|13|21|22|23|31|32|33|61|62].bo3.*