I have a pool of 4 older E5545 servers in a Twin^2 SuperMicro chassis, three of which I have powered down for heat reasons and power usage while swapping out 120VAC UPS' for a 208VAC UPS. After starting to shut down some of the other nodes, and leaving the master node up, I noticed that the stats for the two SR's they shared, a SMB and iSCSI SR, started to no longer have stats in XO despite still obviously communicating with them.
I assume its stuck in syncIOwait to gather the states from the powered down nodes but never recovers from this state.
There is JS console errors as well (Chrome 75, Win10 64, XO 5.43, XCP-ng 7.6):
Uncaught (in promise) t
code: 18
data: {objectId: "c60d7e6f-8cbf-4ffd-9241-9938c3162f38"}message: "host offline"name: "t"stack: "t: host offline at t.<anonymous> (http://xen-orchestra/index.js:1:3602096) at g (http://xen-orchestra/index.js:1:223360) at Generator._invoke (http://xen-orchestra/index.js:1:223148) at Generator.e.<computed> [as next] (http://xen-orchestra/index.js:1:223539) at r (http://xen-orchestra/index.js:1:213429) at http://xen-orchestra/index.js:1:213570 at new Promise (<anonymous>) at new t (http://xen-orchestra/index.js:1:413522) at t.<anonymous> (http://xen-orchestra/index.js:1:213364) at t.<anonymous> (http://xen-orchestra/index.js:1:3602652)"__proto__: i
I'm not sure what I could do to help it other than spin up the other three nodes just to regain the stats (trying to monitor testing servers to spec out final homes).