hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...] On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com wrote: hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm_campaign=sent-from-mailbird
On 22/05/2015 14:11:52, Farshid Abediny <farshidabediny@gmail.com javascript:_e(%7B%7D,'cvml','farshidabediny@gmail.com');> wrote: hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org javascript:_e(%7B%7D,'cvml','observium@observium.org'); http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote:
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system,
we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com
wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com
On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote:
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote:
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system,
we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com
wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
That depends solely on the kind of hardware you use. We've pulled over 1.5 million IOPS with a single KVM instance 4 years ago, I'm sure mysql and rrd require much less.
On Fri, May 22, 2015 at 9:43 AM, Adam Armstrong adama@memetic.org wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com
On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote:
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote:
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system,
we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com
wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
*woosh*
On 2015-05-22 14:45, Dan Yasny wrote:
That depends solely on the kind of hardware you use. We've pulled over 1.5 million IOPS with a single KVM instance 4 years ago, I'm sure mysql and rrd require much less.
On Fri, May 22, 2015 at 9:43 AM, Adam Armstrong adama@memetic.org wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com [1]
On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote: Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote: Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote: We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm... [3]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
Links:
[1] http://www.aqua-mail.com [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
I run our observium on Proxmox OpenVZ container!
2GB ram, 2 cores, 20GB HDD
never uses more than 1.5Gb ram and 22% cpu
thats monitoring 50 devices and 600 ports, and even using syslog integrated with observium too
works brilliant too when i need to do hardware maintaince, i just migrate the container to another server and it just carrys on :)
Simon
On 22 May 2015, at 2:43 pm, Adam Armstrong adama@memetic.org wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com http://www.aqua-mail.com/ On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote:
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong <adama@memetic.org mailto:adama@memetic.org> wrote: Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote: We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong <adama@memetic.org mailto:adama@memetic.org> wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny <farshidabediny@gmail.com mailto:farshidabediny@gmail.com> wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org mailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm... http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm_campaign=sent-from-mailbird
observium mailing list observium@observium.org mailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org mailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Try 150,000 ports, then VM overheads will make you cry :)
adam.
Sent from Mailbird [http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...] On 22/05/2015 14:49:11, Simon Smith simonsmith5521@gmail.com wrote: I run our observium on Proxmox OpenVZ container!
2GB ram, 2 cores, 20GB HDD
never uses more than 1.5Gb ram and 22% cpu
thats monitoring 50 devices and 600 ports, and even using syslog integrated with observium too
works brilliant too when i need to do hardware maintaince, i just migrate the container to another server and it just carrys on :)
Simon
On 22 May 2015, at 2:43 pm, Adam Armstrong <adama@memetic.org [mailto:adama@memetic.org]> wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation. Adam. Sent with AquaMail for Android http://www.aqua-mail.com [http://www.aqua-mail.com/] On 22 May 2015 14:36:25 Dan Yasny <dyasny@gmail.com [mailto:dyasny@gmail.com]> wrote: Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong <adama@memetic.org [mailto:adama@memetic.org]> wrote:
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong <adama@memetic.org [mailto:adama@memetic.org]> wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny <farshidabediny@gmail.com [mailto:farshidabediny@gmail.com]> wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org [mailto:observium@observium.org] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [http://postman.memetic.org/cgi-bin/mailman/listinfo/observium] [1]
Links: ------ [1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [http://postman.memetic.org/cgi-bin/mailman/listinfo/observium] [2] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm... [http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&...]
_______________________________________________ observium mailing list observium@observium.org [mailto:observium@observium.org] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [http://postman.memetic.org/cgi-bin/mailman/listinfo/observium]
_______________________________________________ observium mailing list observium@observium.org [mailto:observium@observium.org] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [http://postman.memetic.org/cgi-bin/mailman/listinfo/observium]
_______________________________________________ observium mailing list observium@observium.org [mailto:observium@observium.org] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [http://postman.memetic.org/cgi-bin/mailman/listinfo/observium]
_______________________________________________ observium mailing list observium@observium.org [mailto:observium@observium.org] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
*woosh* (C)
On Fri, May 22, 2015 at 9:54 AM, Adam Armstrong adama@memetic.org wrote:
Try 150,000 ports, then VM overheads will make you cry :)
adam.
Sent from Mailbird http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm_campaign=sent-from-mailbird
On 22/05/2015 14:49:11, Simon Smith simonsmith5521@gmail.com wrote: I run our observium on Proxmox OpenVZ container!
2GB ram, 2 cores, 20GB HDD
never uses more than 1.5Gb ram and 22% cpu
thats monitoring 50 devices and 600 ports, and even using syslog integrated with observium too
works brilliant too when i need to do hardware maintaince, i just migrate the container to another server and it just carrys on :)
Simon
On 22 May 2015, at 2:43 pm, Adam Armstrong adama@memetic.org wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com
On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote:
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote:
Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote:
We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system,
we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com
wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
No, you both missed the point.
These two pieces of data aren't important:
a) a tiny deployment inside a vm b) the existence of giant vm platforms
The only data which is relevant is that virtualisation has a significant impact on Observium performance on very large installations.
People with very large installations are probably better off not running in VMs because it degrades performance not only for Observium, but anything else using the VM platform. If you have a small installation, it doesn't really matter.
adam.
On 2015-05-22 14:56, Dan Yasny wrote:
*woosh* (C)
On Fri, May 22, 2015 at 9:54 AM, Adam Armstrong adama@memetic.org wrote:
Try 150,000 ports, then VM overheads will make you cry :)
adam.
Sent from Mailbird [1]
On 22/05/2015 14:49:11, Simon Smith simonsmith5521@gmail.com wrote:I run our observium on Proxmox OpenVZ container!
2GB ram, 2 cores, 20GB HDD
never uses more than 1.5Gb ram and 22% cpu
thats monitoring 50 devices and 600 ports, and even using syslog integrated with observium too
works brilliant too when i need to do hardware maintaince, i just migrate the container to another server and it just carrys on :)
Simon
On 22 May 2015, at 2:43 pm, Adam Armstrong adama@memetic.org wrote:
Or that, but virtualization is not a great solution for a heavily loaded observium installation.
Adam.
Sent with AquaMail for Android http://www.aqua-mail.com [2]
On 22 May 2015 14:36:25 Dan Yasny dyasny@gmail.com wrote: Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong adama@memetic.org wrote: Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote: We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong adama@memetic.org wrote:
No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2]
On 22/05/2015 14:11:52, Farshid Abediny farshidabediny@gmail.com wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3] [1]
Links:
[1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3] [2]
http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm... [4]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3]
Links:
[1] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm... [2] http://www.aqua-mail.com/ [3] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [4] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&...
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Am Fr, 22.05.2015 um 14:54 schrieb "Adam Armstrong" adama@memetic.org:
Try 150,000 ports, then VM overheads will make you cry :)
OpenVZ is does not have the VM overhead (same like LXC and docker). It talks directly to the hardware -- makes a huuuuuge difference and feels like a real server.
For kvm and ESX I totally agree. The IO got very much in the way with more than 10000 ports.
/Steffen
I agree, and in vwmare you can run a fault tolerance VM meaning you will always have a second copy.
But otherwise do as a regular application:
· Separate application from database – setup your own DB farm
· Have a redundant file system
We run everything virtualized, 125.000 exchange mailboxes. Virtual video encoders (with shared GPUs), I bet a cent or two that 150.000 ports is less consuming comparted to what else we do, and our Observium only has ~10.000 ports.
From: observium [mailto:observium-bounces@observium.org] On Behalf Of Dan Yasny Sent: Friday, May 22, 2015 3:36 PM To: Observium Network Observation System Subject: Re: [Observium] Observium cloud
Why work so hard? Anything can be made HA when put into VMs. My monitoring server is a VM, if the hardware it's on dies, it simply gets started elsewhere
On Fri, May 22, 2015 at 9:27 AM, Adam Armstrong <adama@memetic.orgmailto:adama@memetic.org> wrote: Quis custodiet ipsos custodes?
You seem to be trying terribly hard to vastly overcomplicate your deployment.
You can certainly run two installations (though that would require two licenses), but they won't communicate in any way, since we don't have any form of replication.
I'd be inclined to just rsync RRDs and copy the database and config periodically to a second server, so you can switch over to a second system if the first one explodes. You'll lose data between the most recent backup and the system failure, but you'll gain not having to spend the £20,000 we'd be charging if we supported full active/active replication.
I don't consider metric collection and monitoring to be a business critical service, so doesn't really require replication and failover. Even losing a few hours of billing data isn't really an issue.
Just have a decent backup/restore procedure and you that's sufficient.
It's also true that a decent backup policy is far safer than relying on active replication, since a software or configuration issue will destroy both of your instances. Just ask people who've run VSS!
adam.
On 2015-05-22 14:17, Farshid Abediny wrote: We want install on 2 node that when our first node goes down the second node can monitor for example node 1 connection lost
On Friday, May 22, 2015, Adam Armstrong <adama@memetic.orgmailto:adama@memetic.org> wrote: No. And since it only takes 15 minutes to install a new system, we're not convinced that the massive increase in complexity (and network load) would be a good idea.
Observium is literally config.php, a MySQL database and a directory of RRDs. Complex stuff isn't really needed, nor particularly helpful.
Just make periodic backups and make sure you know how to restore the MySQL database and RRD directory on failure.
adam.
Sent from Mailbird [2] On 22/05/2015 14:11:52, Farshid Abediny <farshidabediny@gmail.commailto:farshidabediny@gmail.com> wrote:hi I want know can we install observium on 2 server and if one server fail then use our second server for monitoring is it possible? Thanks _______________________________________________ observium mailing list observium@observium.orgmailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [1]
Links: ------ [1] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] http://www.getmailbird.com/?utm_source=Mailbird&utm_medium=email&utm...
_______________________________________________ observium mailing list observium@observium.orgmailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
_______________________________________________ observium mailing list observium@observium.orgmailto:observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
participants (6)
-
Adam Armstrong -
Anders Johansson -
Dan Yasny -
Farshid Abediny -
Simon Smith -
Steffen Klemer