load-balancing. what do I miss ?



  • ok, step by step I will learn where is everything 馃槃

    ok there is what I think is related to the load-balancing
    "FaultyGranularity: Unable to get the true granularity: 5"

    May 30 19:59:00 XO xo-server[405]: [WARN] scheduled function: FaultyGranularity: Unable to get the true granularity: 5
    May 30 19:59:00 XO xo-server[405]:     at /opt/xen-orchestra/packages/xo-server/src/xapi-stats.js:321:12
    May 30 19:59:00 XO xo-server[405]:     at Generator.next (<anonymous>)
    May 30 19:59:00 XO xo-server[405]:     at step (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:221)
    May 30 19:59:00 XO xo-server[405]:     at _next (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:409)
    May 30 19:59:00 XO xo-server[405]:     at run (/opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:75:22)
    May 30 19:59:00 XO xo-server[405]:     at /opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:92:30
    May 30 19:59:00 XO xo-server[405]:     at flush (/opt/xen-orchestra/node_modules/core-js/modules/_microtask.js:18:9)
    May 30 19:59:00 XO xo-server[405]:     at _combinedTickCallback (internal/process/next_tick.js:131:7)
    May 30 19:59:00 XO xo-server[405]:     at process._tickCallback (internal/process/next_tick.js:180:9)
    May 30 19:59:02 XO xo-server[405]: xo-server-cloud: next attempt in 32879 ms
    May 30 19:59:35 XO xo-server[405]: xo-server-cloud: next attempt in 57337 ms
    May 30 19:59:42 XO xo-server[405]: 2018-05-30T17:59:42.422Z xo:perf blocked for 858ms
    May 30 20:00:00 XO xo-server[405]: [load-balancer]Execute plans!
    


  • It's a "from the sources" install right? Why on earth do you have the xo-server-cloud plugin Oo

    Anyway, yes, the load balancer isn't able to fetch the stats in your case, I don't know why.



  • yes it's XOCE.

    well I installed it with an script, and I guess that "xo-server-cloud" was enabled by default, so I disabled it, and tried again, but same error on the 2 pools that I have with 2 XOCEs (one in each pool).

    The XenServer is a 7.1, may be today I'm going to update the dev pool to XCP-ng 7.4.1

    I'll keep you posted.



  • Why having 2 XOs on 2 different pool? It's an anti-pattern. It's one XO to control all your pools.



  • 馃槵 didn't know that. It was there because it was a dev pool and for testing purposes.

    Ok, by the way I stopped one of them, but same errors happens.

    well I will leave the log error here, may be with an update this will be solved.
    At this moment I really don't need it, but I wanted to see it in action.

    May 31 11:29:00 XO xo-server[947]: [WARN] scheduled function: FaultyGranularity: Unable to get the true granularity: 5
    May 31 11:29:00 XO xo-server[947]:     at /opt/xen-orchestra/packages/xo-server/src/xapi-stats.js:321:12
    May 31 11:29:00 XO xo-server[947]:     at Generator.next (<anonymous>)
    May 31 11:29:00 XO xo-server[947]:     at step (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:221)
    May 31 11:29:00 XO xo-server[947]:     at _next (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:409)
    May 31 11:29:00 XO xo-server[947]:     at run (/opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:75:22)
    May 31 11:29:00 XO xo-server[947]:     at /opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:92:30
    May 31 11:29:00 XO xo-server[947]:     at flush (/opt/xen-orchestra/node_modules/core-js/modules/_microtask.js:18:9)
    May 31 11:29:00 XO xo-server[947]:     at _combinedTickCallback (internal/process/next_tick.js:131:7)
    May 31 11:29:00 XO xo-server[947]:     at process._tickCallback (internal/process/next_tick.js:180:9)
    May 31 11:30:00 XO xo-server[947]: [load-balancer]Execute plans!
    


  • Try to grep on load\-balancer



  • ok this is it

    May 31 12:41:00 XO xo-server[947]: [load-balancer]Execute plans!
    May 31 20:47:00 XO xo-server[947]: [load-balancer]Execute plans!
    May 31 20:47:00 XO xo-server[947]: [load-balancer]No hosts to optimize.
    May 31 20:48:00 XO xo-server[947]: [load-balancer]Execute plans!
    May 31 20:48:01 XO xo-server[947]: [load-balancer]No hosts to optimize.
    May 31 20:49:00 XO xo-server[947]: [load-balancer]Execute plans!
    May 31 20:49:00 XO xo-server[947]: [load-balancer]No hosts to optimize.
    May 31 20:50:00 XO xo-server[947]: [load-balancer]Execute plans!
    May 31 20:51:00 XO xo-server[947]: [load-balancer]Execute plans!
    

    and with a normal "tail" of /var/log/syslog

    May 31 20:52:00 XO xo-server[947]: [WARN] scheduled function: FaultyGranularity: Unable to get the true granularity: 5
    May 31 20:52:00 XO xo-server[947]:     at /opt/xen-orchestra/packages/xo-server/src/xapi-stats.js:321:12
    May 31 20:52:00 XO xo-server[947]:     at Generator.next (<anonymous>)
    May 31 20:52:00 XO xo-server[947]:     at step (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:221)
    May 31 20:52:00 XO xo-server[947]:     at _next (/opt/xen-orchestra/packages/xo-server/dist/xapi-stats.js:44:409)
    May 31 20:52:00 XO xo-server[947]:     at run (/opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:75:22)
    May 31 20:52:00 XO xo-server[947]:     at /opt/xen-orchestra/node_modules/core-js/modules/es6.promise.js:92:30
    May 31 20:52:00 XO xo-server[947]:     at flush (/opt/xen-orchestra/node_modules/core-js/modules/_microtask.js:18:9)
    May 31 20:52:00 XO xo-server[947]:     at _combinedTickCallback (internal/process/next_tick.js:131:7)
    May 31 20:52:00 XO xo-server[947]:     at process._tickCallback (internal/process/next_tick.js:180:9)
    

    also there are the snapshots of how its the load-balancing configured and the actual status of the pool
    0_1527792620955_Selecci贸_089.png 0_1527792645216_Selecci贸_090.png



  • hi ! I "reopen" the question again.

    I'm 5.21.0 web and server, there is not error on the log file, but the balancing is not being done.

    First. I set a 20% CPU maximum use of CPU, on a VM I started stress-ng, and the CPU rised up to 100%, and nothing happened.

    The second one I set a maximum of free RAM to 1000MB, I started machines only on that host and it left 500MB free, bunt nothing happened.

    (by the way, I should be more useful to have a % free RAM instead an static number)

    Thanks.

    I tried a different combinations

    20%CPU
    0_1530254901369_Selecci贸_049.png 0_1530254898188_Selecci贸_048.png

    Second test. 1000 MB RAM
    0_1530254939227_Selecci贸_051.png 0_1530254909321_Selecci贸_050.png



  • I am having the exact same issue as this previous thread with Performance mode.
    Second host is not being utilized at all.
    I can cause the optimization message to happen with CPU usage but not with RAM.
    Either way it doesn't move any VMs in Performance mode.

    If i switch to Density it will move all the VMs to one server and shut down the cleared up server.

    Does anyone actually have this working?



  • Hi, this forum isn't used anymore, it was migrated to XCP-ng Forum, in the "Xen Orchestra" section. https://xcp-ng.org/forum

    Feel free to create a topic there 馃檪

    Note: alternatively, you can open a support ticket so a XO dev could take a look remotely on your setup.


Log in to reply