Can someone explain the difference between Virtualization Manager and IVIM? What do I get from Virtualization Manager that I don't get from IVIM? Does it require a separate machine and cannot be installed on the same machine as NPM and all the other modules?
↧
Virtualization Manager Versus IVIM
↧
Vcenter not polling in VMan
Hi,
In our Orion VMan we have enabled vmware polling in some of the vcenters. Of those in one of the Vcenters we are getting the error like " Error while connecting to VMware device - There was no endpoint listening at https://<IPaddress>/sdk that could accept the message. This is often caused by an incorrect address or SOAP action. See InnerException, if present, for more details."
When we try to test the credentials in the Poll for VMware in the Add node page getting the error as "Test Failed. ScheduleVMwareCredentialsTestProbeJobs failed, check fault information." The user which I am trying to test is having admin privilege in the vcenter.
Can some one help me to resolve this.
Regards,
Sreenath
↧
↧
Email Notifications
I am trying to set up email notifications for certain things such as High CPU usage, memory usage, etc. I can't seem to get the email to send when this happens, where do I set this up to make this work?
↧
How to alert on vMan Component Volumes?
Environment: vMan Orion 8.1 (no appliance) and Orion 2017.3 with SAM 6.5
Objective: Alert on disk utilization from any VMs managed by vMan. These VMs are only managed by vMan and not managed by SAM thus the VMs are not managed nodes from Orion perspective. In other words, I like to alert when vMan VM Component Volumes object shows in Red:
I don't see any option to specify these objects and triggers when building alerts.
Thanks
j
↧
Monitoring RAID Status for ESXi Servers
Hi!
Is VMan capable of monitoring the RAID Array status of ESXi servers? I specifically mean the direct-attached disks that are RAID-configured on a VMWare server.
Can't find anything in the VMan guide. Hope somebody can help.
Thanks in advance,
Paulo
↧
↧
VM Disk Latency Alerts
Since upgrading our environment to ESXi/vCenter 6.5U2 we have been experiencing "VM Disk Latency" according to VMAN. Frequently it will show latency approaching 25 seconds! After hours of trying to find the cause of this latency we are starting to wonder if the data in Orion could be incorrect. We can not find any evidence of latency exceeding 8ms in vCenter nor on the hosts using esxtop.
Where is VMAN pulling the VM Disk stats from and how can I validate that an issue actually exists?
*Note: VMAN 8.2.1 HF1
↧
VMware 6.5 and high VM latency
We just upgraded to our vCenter and Hosts to VMware 6.5 update 1. After the upgrade we are seeing high vm latency in vman. Datastore latency seems to be reporting correctly, but vm latency is reporting things like 100,000y. If you look at the vm performance within vCenter and or ESXi top we are not seeing those latency's and performance is as expected. We are running vman 7.1 which is supposed to be supported in 6.5. Was wondering if anyone has seen this.
↧
VMan - Cluster low VM capacity - falsely shows 0 VMs left
The alert in my VMan instance (v4.1.0.10002) for "Cluster low VM capacity" stays triggered and thinks that I have no capacity left for VMs due to "disk IO". This is the case for both of my vSphere clusters (but it isn't true).
When I look at my cluster details, under the Performance section, "Avg. IOPS" has a value of "--- cmd/s" (not sure if that should have a value or is relevant). I only mention that because in times past, we've had VMan bugs and false positives due to VMan retrieving null values.
Anyone have an idea of how to fix this? Just wanted to ping out here before hitting up support.
Thanks.
↧
"vmStatsProvider" causing Solarwinds to give false "down" alerts?
Hi Everyone,
Does anyone have experience Solarwinds giving false "down" alerts on with virtual servers?
It happened on a server yesterday morning, twice. We received alerts from Solarwinds that a particular server was down around 7:00 and 8:00 AM.
In looking at the application logs for that time on the server I spied numerous line items:
| Information | 10/9/2014 7:08 | vmStatsProvider | 256 | General | The "vmStatsProvider" is successfully initialized for this Virtual Machine. WMI namespace: "root\cimv2". | |||||||||
| Information | 10/9/2014 7:08 | vmStatsProvider | 258 | Guest Library API | The "vmGuestLibrary" is successfully initialized for this Virtual Machine. | |||||||||
| Information | 10/9/2014 7:06 | vmStatsProvider | 256 | General | The "vmStatsProvider" is successfully initialized for this Virtual Machine. WMI namespace: "root\cimv2". | |||||||||
| Information | 10/9/2014 7:06 | vmStatsProvider | 258 | Guest Library API | The "vmGuestLibrary" is successfully initialized for this Virtual Machine. | |||||||||
| Information | 10/9/2014 6:57 | vmStatsProvider | 256 | General | The "vmStatsProvider" is successfully initialized for this Virtual Machine. WMI namespace: "root\cimv2". | |||||||||
| Information | 10/9/2014 6:57 | vmStatsProvider | 258 | Guest Library API | The "vmGuestLibrary" is successfully initialized for this Virtual Machine. | |||||||||
| Information | 10/9/2014 6:55 | vmStatsProvider | 256 | General | The "vmStatsProvider" is successfully initialized for this Virtual Machine. WMI namespace: "root\cimv2". | |||||||||
| Information | 10/9/2014 6:55 | vmStatsProvider | 258 | Guest Library API | The "vmGuestLibrary" is successfully initialized for this Virtual Machine. | |||||||||
| Information | 10/9/2014 6:46 | vmStatsProvider | 256 | General | The "vmStatsProvider" is successfully initialized for this Virtual Machine. WMI namespace: "root\cimv2". | |||||||||
| Information | 10/9/2014 6:46 | vmStatsProvider | 258 | Guest Library API | The "vmGuestLibrary" is successfully initialized for this Virtual Machine. | |||||||||
If I understand this correctly these are coming from the VMWare tools that is installed on the virtual instance?
Can this be causing Solarwinds to see the server being down?
Thanks everyone,
Cheryl
↧
↧
High Percent Co-Stop (%CSTP)
Co-Stop (%CSTP) is the percentage of time that a SMP virtual machine was ready to run, but incurred delay due to co-vCPU scheduling contention. To remedy Co-Stop on a VM, decrease the number of vCPUs on that VM. If there are multiple VMs with high %CSTP on a host, it is likely a sign of CPU contention and resource starvation due too many vSMP VMs on that host. Migrate some of those VMs off to other hosts to re-balance the load in the cluster and also, decrease the number of vCPUs allocated to those VMs to see if the %CSTP goes to zero.
| VM CPU Co-stop (%CSTP) Alert in Virtualization Manager | VM CPU Co-Stop Alert Viewer in Virtualization Manager |
|---|---|
|  |
↧
VIM Max Iops vs VMAN Peak Iops
I'm trying to run some reports to see what datastores have been exceeding a few different thresholds. When I pull datastore statistics in orion, maxiops never seems to get too high. However charting a few datastores in VMAN that i know are heavy hitters i see some pretty substantial peaks.
How are these calculated differently? Also, in vman is it possible to run reports against the historical data? I know that data points are there, just not sure how to run reports against it and export.
Thanks.
↧
How is hardware status data propagated from VMAN to Orion via Integration?
With Virtualization Manager enabled, we are sending data from our VMAN environment into our Orion instance. The datacenter engineers reached out because they had replaced some failed hardware in a host in our cluster and asked how long it took for that data to update.
When I checked the status, I saw that the last polled time was 4:26AM. That's an odd time since I know that the Hardware Details statuses come from the VMware API when VMAN and Orion are integrated. We are using the default 12 hour interval for configuration polling and, when I checked, the configuration for the vCenter that manages this host was being polled at 10:30ish. I kicked off a configuration poll anyway. I jumped into the VIM_Hosts table and searched for my host to get the VIM_Hosts.ManagedObjectID (host-178515 in my case). Then I jumped out to my VMAN appliance and did a 'cat /var/hyper9/logs/master.log | grep -i host-178515'
2017-01-13 10:24:52,680 [ConfigData-8] DEBUG com.hyper9.service.configitem.storage.DataStorageProcessor$ConfigDataTask:517 - Processed config data for HostServer host-178515 for job aedecbb8-22dc-4df6-aa76-471aec556580 in 1058.771 ms
I found this entry which indicated that the config data (vs the dozens of other log entries that said 'Processed 5120 samples' -- that is the performance data) but still no change in the hardware status on that node.
I dug a little deeper and looked up the node in APM_HardwareInfo by the nodeID (also found in the VIM_Hosts table) and found that the LastPollTime was 2016-11-13 09:26:43.330 (NOTE: This is obviously UTC time even though the field name doesn't say so!) and that the categories with problems shows the piece of hardware that has an issue.
Now I am *really* confused because I thought that I should have been able to force an update to this node from VMAN by pulling the configuration data.
After all that my question is -- How is the hardware status data updated in Orion if you have a VMAN instance integrated? How long should it take to refresh that status data after a change is made?
↧
Report for Average Memory and CPU Utilization
I am relatively new to VMAN. Is there a way to create a tabular report showing a bunch of different systems with columns for Average Memory and CPU utilization over a period of time?
↧
↧
Datastore low free space alert
Will VMAN trigger a datastore low free space alert if the VMs attached to the datastore are allocated over 90% of datastore capacity using thick disk or is the alert triggered when actual stored data (OS, apps, files, etc) reach over 90%
↧
VMware Polling / SNMPv3
Hi All,
I thougnt I'd post the question here in case anyone can help or has the same query.
The site I am working on is looking at VMAN, NPM and SAM as a base product set. This particular query is around VMware vCenter as well as ESXi polling.
I have added a vCenter Appliance as a monitored Node using Status Only: ICMP and Poll for VMware Providing creds.
When I clicked through the vCenter Node to get to an ESXi host, it was unmanaged to I added this in the same way. The wizard stated that I was polling the ESXi host via the vCenter and that I should use Status Only: ICMP as the polling method.
From what I understand, if I want to view VM interfaces stats (i.e. virtual NICS) using an Agent this does not count against my NPM interface count, which is great.
What I'm unclear about is how I monitor network interfaces on the ESXi hosts themselves. From what I can see, this isn't done via vCenter so I assume I would need to enable SNMP in order to get this information, which goes against the Add Node wizard guidance.
On this point also, if I want the vCenter to be able to send SNMPv3 traps to Orion, presumably I need to add it as an SNMPv3 managed host?
Unless I'm missing something, I can't see how ESXi interface statistics get monitored via any other means than SNMP.
Has anyone got any more info on this that might help?
Many thanks!
M
↧
Is multi-hypervisor support a must-have in management tools?
Hello, fellow thwackers!
Let me ask you a question: When you look at buying virtualization management tools, whether you’re looking for tools to perform comprehensive management of the virtual infrastructure or you’re on the hunt for a robust health, monitoring and capacity management tool, how important is it to you that your tools of choice be hypervisor agnostic?
You may wonder why this would be important since the vast majority of virtual implementations happen on VMware and all of the major third party players provide VMware support first and, maybe, other hypervisors.
Here’s why: I predict that, within the next 3 to 5 years, Hyper-V will eat VMware’s lunch. VMware’s share of the market will plummet by double-digit percentages as Hyper-V 3.0 – and all of its goodies – makes its way into the mainstream. Thus, the ability to support a heterogeneous hypervisor environment will be critical.
Now, before I get flamed, let me explain my reasoning with the following points:
- At present, vSphere is far superior to Hyper-V, although I believe that Hyper-V would fit far more use cases than are in production. With Hyper-V 3.0, Microsoft will go a long way toward shrinking the feature gap that exists between the two product and many more companies will consider Hyper-V adequate for their needs.
- VMware has made some significant missteps and reduced the amount of trust that people have in the company. Two key factors: The Enterprise Plus SKU that was released with vSphere 4 and vRAM-based licensing introduced with vSphere 5. Both “enhancements” provide VMware to extract dollars from existing customers beyond normal annual maintenance. I believe that both were nothing more than money grabs. While many also don’t trust Microsoft, Hyper-V licensing is far less complex.
Before you think I hate VMware, think again. It’s a great product and I’ve run it in multiple environments. However, as time goes on, I see the company facing greater challenges. So, are you preparing for the day when VMware may no longer reign supreme?
↧
vsphere 6 plugin
Sorry for my ignorance. We just upgraded from ESXi 5.1 to 6.0 and vsphere 5.1 windows to VCSA 6.0 u1. Can someone remind me how do install the Solarwinds Plugin into the environment?
↧
↧
Number of hosts per cluster in Report?
I am trying to include the number of hosts in each VMware cluster in my VMAN Orion reports, but cannot seem to find it in the available data. VMAN definitely knows the number of hosts per cluster, because if you hover over a cluster it will show #Hosts. What am I missing? Thanks.
↧
Vcenter not polling in VMan
Hi,
In our Orion VMan we have enabled vmware polling in some of the vcenters. Of those in one of the Vcenters we are getting the error like " Error while connecting to VMware device - There was no endpoint listening at https://<IPaddress>/sdk that could accept the message. This is often caused by an incorrect address or SOAP action. See InnerException, if present, for more details."
When we try to test the credentials in the Poll for VMware in the Add node page getting the error as "Test Failed. ScheduleVMwareCredentialsTestProbeJobs failed, check fault information." The user which I am trying to test is having admin privilege in the vcenter.
Can some one help me to resolve this.
Regards,
Sreenath
↧
How to Bulk Add ESX Hosts as Orion Nodes in new Orion VMAN 8.2?
Either SolarWInds completely overlooked this when changing how Virtualization Objects are added to Orion, or I am completely missing the setting to do this for us.
Previously you could just run a one hop Network Discovery with the vCenter server as the only IP address. You just had to put in valid VMware credentials and Everything would get brought in as Orion nodes and with the VMAN appliance integrated it would automatically marry the data up. You could even put in WMI and SNMP credentials on the appropriate pages of the Discovery wizard and it would even scan and bring in all of the VM's as well. It worked great for bulk adding all of our ESX hosts and VM's into Orion in one big swoop.
Now however, the page in the Discovery wizard where you used to put in the VMware credentials has been replaced by the "Add VMWare VCenter or Hyper-V Devices" button. Clicking this button exits you from the Discovery Settings wizard and takes you to the new Add Virtualization Objects page. I thought to myself, "OK, cool, so we just add the vCenter server and it will now automatically add the ESX Hosts in for us. I hope it also adds in the VM's like the old Discovery Wizard does too, but as long as it at least gets the Hosts I'll be happy.". I figured surely they wouldn't overlook that, right?
Well, they did... Adding the vCenter server this way only brings the vCenter server in as an Orion node. The ESX hosts are only brought into Orion as "Orion VMAN" nodes. This means the hosts are not pinged by Orion every 2 minutes, they are not polled for additional metrics like CPU, SNMP, HARDWARE HEATLH, or any of the other data that Orion gathers. It also means you can't assign custom property values to them or anything. This is a HUGE miss by SolarWinds. We have nearly 300 ESX hosts between our two vCenters. Clicking on each one individually and choosing "Manage This Node" is just not feasible.
The only other thing I can think to do would be to do a SWQL\SQL query to get a list of all the ESX Host IP's from the VIM.Hosts table, then run a "Ping Only" discovery against that list of IP's. It should marry them up once brought in (I hope). This is a pain though and is a big step backwards compared to how it used to work.
Luckily I had two Discovery Profiles saved that are setup the way I described above. Running them worked just like it used to and brought in all the VMware Hosts and VM's as both Orion and Orion VMAN nodes. But what if these profiles ever get deleted? What if we get a new vCenter server that I need to Discover?
Please SolarWinds, tell me you have a function built somewhere that replaces the features I listed above? We are a large environment, so when you remove functionality that allowed us to bulk manage things in Orion that really hampers us and makes our job harder. What is SolarWinds vision for Orion VMAN with regards to monitoring hosts? Do you intend for your users to just have to click on each host individually to manage them as Orion Nodes? Do you intend for us to only manage them as VMAN Nodes and thus no longer get hardware health or other important data that is collected by Orion core? I just don't understand how such a major thing was overlooked? I scoured the admin guides and documentation last night and could not find a single word about adding hosts in (other than adding standalone esx hosts). Our use case is not unique by any stretch of the imagination, so how are other users handling this? I really don't get it at all. It is baffling to me how this oversight exists... Or, like I said, maybe I'm just missing something. I sincerely hope that is the case.
↧