Hey, just FYI.
Hypervisor: Xenserver 6.5 / latest Patches VM: From Template "[x] Dynamic scalable" OS Type [Centos] - Inside a cloud-prepared Centos 7.0 plain. Offering: 4GB RAM (XenServer itself sets dynamic-min and dynamic-max to 4G, static-max to around 12G - which is intentional) As of Centos 7.0 running, the VM shows: [root@prometheus ~]# free -g total used free shared buff/cache available Mem: 3 0 1 0 1 2 Swap: 1 0 1 After a plain "yum update -y && reboot", this changes to: root@prometheus ~]# free -g total used free shared buff/cache available Mem: 12 0 12 0 0 12 Swap: 1 0 1 So now, the machine looks like it could consume up to 12 G... well, the mtrr maxes out at 4G. If one tries to consume more than the original 4G, XenServer (and subsequently Cloudstack) looses the machine to a Xen OOM. It shuts down (the hard way). To mitigate that, powering off, changing the OS Type to Centos 7.2 and starting up again fixes that issue. So again: vmadmin@prometheus ~]$ free -g total used free shared buff/cache available Mem: 3 0 3 0 0 3 Swap: 1 0 1 We double checked that issue on two diferent clusters with the very same result, but didn't dug too deep into kernelspace to find a reason. By now, we're happy to solve that by just changing the OS Type. Mit freundlichen Grüßen, Stephan Seitz -- Heinlein Support GmbH Schwedter Str. 8/9b, 10119 Berlin http://www.heinlein-support.de Tel: 030 / 405051-44 Fax: 030 / 405051-19 Zwangsangaben lt. §35a GmbHG: HRB 93818 B / Amtsgericht Berlin-Charlottenburg, Geschäftsführer: Peer Heinlein -- Sitz: Berlin
signature.asc
Description: This is a digitally signed message part
