Feedback in regards to upcoming XO update
-
Re: Simplifying IT: 2CRSi's Mona Servers with XCP-ng Onboard
@MathieuRA I today noticed this Pull Request (https://github.com/vatesfr/xen-orchestra/pull/7838), however it just mentions the current and/or highest temperature readings. When managing servers its also important to consider the coldest temperatures as well, due to electrical equipment having a lowest specified temperature in the manual.
You wouldn't want conductive liquid vapours condensing or freezing inside an operational server now would we. Liquids such as water and electricity don't mix - short circuit, damage, electrical fire and/or electrocution!
Also the lowest minimum temperature is also likely to be checked as part of a warranty claim investigation (if a server is sent back to manufacturer, or check by manufacturer's agent).
So having this information easily accessible is important! Since there may not be an equivalent to HPE iLO or Dell iDRAC on these server's it may be worth adding the ability for XOA to send alerts based on some alarm values which can be configured!
-
You are reading a PR without context. We are NOT displaying the highest temp in history, but the highest current temp of the 2x CPUs, that's it. There's no reason to have a big diff between the 2, and if yes, then it won't be because it's too cold.
-
@olivierlambert said in Feedback in regards to upcoming XO update:
You are reading a PR without context. We are NOT displaying the highest temp in history, but the highest current temp of the 2x CPUs, that's it. There's no reason to have a big diff between the 2, and if yes, then it won't be because it's too cold.
All right its the current highest temperature of the 2 CPUs, though historical temperatures access of the 2CRSi servers through XOA would a useful QoL feature combined with alerting.
-
That's another topic, but we'll get there at some point
-
@olivierlambert said in Feedback in regards to upcoming XO update:
That's another topic, but we'll get there at some point
Another good addition for this PR would be the current lowest temperature of the 2 CPUs. That way you have context for the current highest temperatures for the 2 CPUs of those servers.
-
If you do not have a 2CRSi machine, you won't have any info at the moment anyway
-
@AtaxyaNetwork Would this additional feedback benefit you if included on the initial plugin release?
-
@john-c The plugin I wrote send all data available to XO, and XO is doing the parsing and display.
As we only work currently with the Mona model, we are not concerned by liquid cooling and other things like this.
To have some history, I recommend using the Netdata plugin and plug it into a Prometheus to have long-term data. For now, XO is displaying real-time value directly from the BMC, with no history.
If you have some values you think are nice to have (in fact, the PSU status was not my idea, but feedback from my community), don't hesitate to tell us !
It's still a work in progress, nothing is set in stone -
@AtaxyaNetwork said in Feedback in regards to upcoming XO update:
@john-c The plugin I wrote send all data available to XO, and XO is doing the parsing and display.
As we only work currently with the Mona model, we are not concerned by liquid cooling and other things like this.
To have some history, I recommend using the Netdata plugin and plug it into a Prometheus to have long-term data. For now, XO is displaying real-time value directly from the BMC, with no history.
If you have some values you think are nice to have (in fact, the PSU status was not my idea, but feedback from my community), don't hesitate to tell us !
It's still a work in progress, nothing is set in stoneThough I wasn't referring to the liquid cooling 2CRSi servers when I wrote the original post. But to the environment where the server is housed, as the conditions of the housing location has an effect on the performance and general health of the server. Hotter environments make the server more unstable (crash more frequently and/or run more slowly); on top of this not last as long due to component damage due to the heat.
On the other end of the scale cooler environments mean the server can run faster, but too cold will enable condensing or freezing of water vapour in the server. This environmental risk can result in damage to servers as a result of the short circuits, these short circuits can also result in fires or electrocutions of people working on the server in the environment of the server.
-
I'm pretty sure this will be caught by XO as soon we manage to parse the values and compare them with the vendor values, which is planned and why we decided to get a partnership with a hardware vendor in the first place: knowing exactly what values are OK and not OK to display that in XO. So that's coming in the future
-
@olivierlambert said in Feedback in regards to upcoming XO update:
I'm pretty sure this will be caught by XO as soon we manage to parse the values and compare them with the vendor values, which is planned and why we decided to get a partnership with a hardware vendor in the first place: knowing exactly what values are OK and not OK to display that in XO. So that's coming in the future
Another vendor of server hardware to approach is in addition to Dell, HPE and 2CRSi the company Novatech. You may say that the are eCommerce selling other people's products but you would be mistaken as they also produce their own servers.
https://www.novatech.co.uk/server/rackmount-server/amd-epyc/
-
I'm not sure to get what you said. Right now, we have a specific partnership with 2CRSi, having access to their engineers who design the case and integrate the components inside. This helps us to give exclusive insight on the hardware, and on their side, providing XCP-ng pre-installed.
I'm not against having other partners, but you can imagine that Dell or HPE are taking more time to provide XCP-ng pre-installed on it. I never heard of NovaTech before, I'll be happy to work with them if they want to work with us
-
@olivierlambert said in Feedback in regards to upcoming XO update:
I'm not sure to get what you said. Right now, we have a specific partnership with 2CRSi, having access to their engineers who design the case and integrate the components inside. This helps us to give exclusive insight on the hardware, and on their side, providing XCP-ng pre-installed.
I'm not against having other partners, but you can imagine that Dell or HPE are taking more time to provide XCP-ng pre-installed on it. I never heard of NovaTech before, I'll be happy to work with them if they want to work with us
I have dropped a message to Novatech with a link to your website letting them know about you. They're currently closed for today but when their people come in tomorrow they'll get my message, be ready during that day onwards for possible contact from them!
Your already got a French municipality as a customer providing their servers with your Virtualisation software stack. Well if you manage to partner with Novatech you'll also have the capacity to gain MBDA along with Thales as customers through them!
They're a solutions producer and provider partnering with them would allow both of you to benefit, as well as incorporate XCP-ng, XOA and/or XOSTOR into the Novatech solutions. Each of the solutions provided by Novatech is bespoke to a specific customer, built using off the shelf commercial components.
Pay attention to their video "Our Story #3 About Us" as they provide worry free computing to the companies working to change the world we live in. All of their people enable their customers to do remarkable things. What's more worry free than not having to worry about, a business's product policy of their closed source proprietary software?
Partnering with them would enable their products and solutions to have Vates VMS as a component to them and reduce the worries of the Novatech customers!
