how to monitor switches and increase the polling time
Hi, since we are monitoring ober 300 servers at a time in a custer, we want to know and analyser the network traffic in ethernet and infiniband switch .how we can monitor it using xymon system monitoring.os is rhel 5.2 and xymon version is 4.2.2 Practical problem happening is that some of the nodes reboots but xymon doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia
-- Pankaj V. Dorlikar
I monitor our switches, and routers using devmon. You then can get all interesting data via SNMP, which works very well. You then can see traffic, vlan settings, error and much more about your network-hardware. It works like a charm in our Xymon 4.2.3.
Maik
On 05/16/2010 05:14 AM, pankaj dorlikar wrote:
Hi, since we are monitoring ober 300 servers at a time in a custer, we want to know and analyser the network traffic in ethernet and infiniband switch .how we can monitor it using xymon system monitoring.os is rhel 5.2 and xymon version is 4.2.2 Practical problem happening is that some of the nodes reboots but xymon doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia
-- Pankaj V. Dorlikar
thanks maik...
On Sun, May 16, 2010 at 5:11 AM, Maik Heinelt <maik at vegasystems.com> wrote:
I monitor our switches, and routers using devmon. You then can get all interesting data via SNMP, which works very well. You then can see traffic, vlan settings, error and much more about your network-hardware. It works like a charm in our Xymon 4.2.3.
Maik
On 05/16/2010 05:14 AM, pankaj dorlikar wrote:
Hi,
- since we are monitoring ober 300 servers at a time in a custer, we want to know and analyser the network traffic in ethernet and infiniband switch .how we can monitor it using xymon system monitoring.os is rhel 5.2 and xymon version is 4.2.2
2)Practical problem happening is that some of the nodes reboots but xymon
doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia
-- Pankaj V. Dorlikar
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
-- Pankaj V. Dorlikar
hi, what can be done for 2 nd query?
---------- Forwarded message ---------- From: pankaj dorlikar <pankaj.dorlikar at gmail.com> Date: Sun, May 16, 2010 at 3:27 PM Subject: Re: [hobbit] how to monitor switches and increase the polling time To: hobbit at hswn.dk
thanks maik...
On Sun, May 16, 2010 at 5:11 AM, Maik Heinelt <maik at vegasystems.com> wrote:
I monitor our switches, and routers using devmon. You then can get all interesting data via SNMP, which works very well. You then can see traffic, vlan settings, error and much more about your network-hardware. It works like a charm in our Xymon 4.2.3.
Maik
On 05/16/2010 05:14 AM, pankaj dorlikar wrote:
Hi,
- since we are monitoring ober 300 servers at a time in a custer, we want to know and analyser the network traffic in ethernet and infiniband switch .how we can monitor it using xymon system monitoring.os is rhel 5.2 and xymon version is 4.2.2
2)Practical problem happening is that some of the nodes reboots but xymon
doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia
-- Pankaj V. Dorlikar
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
-- Pankaj V. Dorlikar
-- Pankaj V. Dorlikar
On Sun, May 16, 2010 07:12, pankaj dorlikar wrote:
hi, what can be done for 2 nd query?
2)Practical problem happening is that some of the nodes reboots but xymon
doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia
Are the nodes in this case servers? The hobbit/xymon client will turn the cpu test yellow for an hour after a reboot. Are you not running the client? You can tweak things to shorten the polling period and do tests in parallel, but you'll probably quickly hit scaling issues on the Xymon server.
2nd case... Well, even if a server reboot between the polling time, Xymon should notice it. In case of a Linux/Unix server, the uptime, which you can see at the CPU data will change and will show you an alert. So increasing the polling time, is not really necessary, I think.
If you wanna get noticed of the behavior, you might should add an alert, to get an email/SMS, if a machine rebooted.
Maik
On 2010/05/16 20:12, pankaj dorlikar wrote:
hi, what can be done for 2 nd query?
---------- Forwarded message ---------- From: *pankaj dorlikar* <pankaj.dorlikar at gmail.com <mailto:pankaj.dorlikar at gmail.com>> Date: Sun, May 16, 2010 at 3:27 PM Subject: Re: [hobbit] how to monitor switches and increase the polling time To: hobbit at hswn.dk <mailto:hobbit at hswn.dk>
thanks maik...
On Sun, May 16, 2010 at 5:11 AM, Maik Heinelt <maik at vegasystems.com <mailto:maik at vegasystems.com>> wrote:
I monitor our switches, and routers using devmon. You then can get all interesting data via SNMP, which works very well. You then can see traffic, vlan settings, error and much more about your network-hardware. It works like a charm in our Xymon 4.2.3. Maik On 05/16/2010 05:14 AM, pankaj dorlikar wrote: Hi, 1) since we are monitoring ober 300 servers at a time in a custer, we want to know and analyser the network traffic in ethernet and infiniband switch .how we can monitor it using xymon system monitoring.os is rhel 5.2 and xymon version is 4.2.2 2)Practical problem happening is that some of the nodes reboots but xymon doesnt show it. it may be because that happens in betwwen the polling period of the server. how to overcome this issue?same thing happens for disk and ncpu and external tests also that it does not identify and show these problems as it must be occuring in betwwn polling period .How to make it fast as ganglia -- Pankaj V. Dorlikar To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk <mailto:hobbit-unsubscribe at hswn.dk>-- Pankaj V. Dorlikar
-- Pankaj V. Dorlikar
participants (3)
-
hobbit＠epperson.homelinux.net
-
maik＠vegasystems.com
-
pankaj.dorlikar＠gmail.com