Monitoring hosts behind a load balancer
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Thanks, Steve Holmes ITaP/Purdue University
We do similar things, but we also monitor each host individually, including the service and ping separately.
---Eric
From: sholmes42 at gmail.com [mailto:sholmes42 at gmail.com] On Behalf Of Steve Holmes Sent: Tuesday, June 15, 2010 12:07 PM To: hobbit at hswn.dk Subject: [hobbit] Monitoring hosts behind a load balancer
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Thanks, Steve Holmes ITaP/Purdue University
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Steve Holmes wrote:
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Have you considered devmon which monitors the f5 itself? I'm not sure what parameters it shows, however. I believe you can also do a content test (eg. your f5 can be configured to say something foreboding and you can configure Xymon to look for it as bad news).
- ---- _ _ _ _ ___ _ _ _ |Y#| | | |\/| | \ |\ | | |Ryan Novosielski - Sr. Systems Programmer |$&| |__| | | |__/ | \| _| |novosirj at umdnj.edu - 973/972.0922 (2-0922) \__/ Univ. of Med. and Dent.|IST/CST-Academic Svcs. - ADMC 450, Newark -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iEYEARECAAYFAkwXp7wACgkQmb+gadEcsb6j6ACfbmXK/tdbUCj+jvDkBr8hUape vjUAn0Cv3AxjT6vPDbe7K3ZuFhYeJ/C0 =GJGN -----END PGP SIGNATURE-----
On Tue, Jun 15, 2010 at 12:18 PM, Ryan Novosielski <novosirj at umdnj.edu>wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Steve Holmes wrote:
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Have you considered devmon which monitors the f5 itself? I'm not sure what parameters it shows, however. I believe you can also do a content test (eg. your f5 can be configured to say something foreboding and you can configure Xymon to look for it as bad news).
No, I had not considered that. I will look into what devmon can do for me. Thanks, Steve
-- The test of a democracy is not the magnificence of buildings or the speed of automobiles or the efficiency of air transportation, but rather the care given to the welfare of all the people. -Helen Adams Keller, lecturer and author (1880-1968)
Truth never damages a cause that is just. -Mohandas Karamchand Gandhi (1869-1948)
I used to have some scripts that would make SOAP requests to an F5 to get the pools and pool member status. I no longer have access to that system, but I may have a backup at home.
I based my scripts on a bunch of example programs in the F5 SDK. IIRC, I altered a couple of the perl scripts to output lines like:
pool1 server1 OK
pool1 server2 DOWN
etc...Then you just run around a loop reading the lines and sending BB status messages.
As long as your F5 is checking the servers occasionally, this approach works OK.
Ralph Mitchell
On Tue, Jun 15, 2010 at 12:07 PM, Steve Holmes <sholmes42 at mac.com> wrote:
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Thanks, Steve Holmes ITaP/Purdue University
On Tue, Jun 15, 2010 at 12:42 PM, Ralph Mitchell <ralphmitchell at gmail.com>wrote:
I used to have some scripts that would make SOAP requests to an F5 to get the pools and pool member status. I no longer have access to that system, but I may have a backup at home.
I based my scripts on a bunch of example programs in the F5 SDK. IIRC, I altered a couple of the perl scripts to output lines like:
pool1 server1 OK pool1 server2 DOWN etc...Then you just run around a loop reading the lines and sending BB status messages.
As long as your F5 is checking the servers occasionally, this approach works OK.
Ralph Mitchell
On Tue, Jun 15, 2010 at 12:07 PM, Steve Holmes <sholmes42 at mac.com> wrote:
Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens.
Thanks, Steve Holmes ITaP/Purdue University
Thanks! I'll check with the F5 admin to see if this might be doable. Steve
-- The test of a democracy is not the magnificence of buildings or the speed of automobiles or the efficiency of air transportation, but rather the care given to the welfare of all the people. -Helen Adams Keller, lecturer and author (1880-1968)
Truth never damages a cause that is just. -Mohandas Karamchand Gandhi (1869-1948)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Steve Holmes wrote:
On Tue, Jun 15, 2010 at 12:42 PM, Ralph Mitchell <ralphmitchell at gmail.com <mailto:ralphmitchell at gmail.com>> wrote:
I used to have some scripts that would make SOAP requests to an F5 to get the pools and pool member status. I no longer have access to that system, but I may have a backup at home. I based my scripts on a bunch of example programs in the F5 SDK. IIRC, I altered a couple of the perl scripts to output lines like: pool1 server1 OK pool1 server2 DOWN etc... Then you just run around a loop reading the lines and sending BB status messages. As long as your F5 is checking the servers occasionally, this approach works OK. Ralph Mitchell On Tue, Jun 15, 2010 at 12:07 PM, Steve Holmes <sholmes42 at mac.com <mailto:sholmes42 at mac.com>> wrote: Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens. Thanks, Steve Holmes ITaP/Purdue UniversityThanks! I'll check with the F5 admin to see if this might be doable. Steve
It would not surprise me though if this is exactly what the F5 devmon template gives you via SNMP. Might be worth a check. If you just download the tar and look at the templates directory under F5, you'll the OID's file should tell you what's monitored. Hate to see duplication of effort. :)
- ---- _ _ _ _ ___ _ _ _ |Y#| | | |\/| | \ |\ | | |Ryan Novosielski - Sr. Systems Programmer |$&| |__| | | |__/ | \| _| |novosirj at umdnj.edu - 973/972.0922 (2-0922) \__/ Univ. of Med. and Dent.|IST/CST-Academic Svcs. - ADMC 450, Newark -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iEYEARECAAYFAkwXsn8ACgkQmb+gadEcsb5+KwCgv2/6zJNTnSQofAK8Hv/goFcs pqgAoJAE38TUulVbW6rwmVbS+lMAz3/1 =AMRr -----END PGP SIGNATURE-----
On Tue, Jun 15, 2010 at 1:03 PM, Ryan Novosielski <novosirj at umdnj.edu>wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Steve Holmes wrote:
On Tue, Jun 15, 2010 at 12:42 PM, Ralph Mitchell <ralphmitchell at gmail.com <mailto:ralphmitchell at gmail.com>> wrote:
I used to have some scripts that would make SOAP requests to an F5 to get the pools and pool member status. I no longer have access to that system, but I may have a backup at home. I based my scripts on a bunch of example programs in the F5 SDK. IIRC, I altered a couple of the perl scripts to output lines like: pool1 server1 OK pool1 server2 DOWN etc... Then you just run around a loop reading the lines and sending BB status messages. As long as your F5 is checking the servers occasionally, this approach works OK. Ralph Mitchell On Tue, Jun 15, 2010 at 12:07 PM, Steve Holmes <sholmes42 at mac.com <mailto:sholmes42 at mac.com>> wrote: Greetings, I'm monitoring several Solaris 10 servers which are behind an F5 load balancer. One of the features of the F5 is that it answers pings for all of the hosts behind it, even if they are all down. Has anyone devised a method of testing the servers for being alive in this context? There are no other network based tests being done. Yes, all of the other tests eventually go purple in Xymon, but we'd like to know that the hosts are down when that happens. Thanks, Steve Holmes ITaP/Purdue UniversityThanks! I'll check with the F5 admin to see if this might be doable. Steve
It would not surprise me though if this is exactly what the F5 devmon template gives you via SNMP. Might be worth a check. If you just download the tar and look at the templates directory under F5, you'll the OID's file should tell you what's monitored. Hate to see duplication of effort. :)
You may well be right. I couldn't tell, back then, because my ex-employer had a habit of blocking SNMP traffic. That might have changed now that they're under new ownership... :)
Ralph Mitchell
participants (4)
-
etmsys＠rit.edu
-
novosirj＠umdnj.edu
-
ralphmitchell＠gmail.com
-
sholmes42＠mac.com