Search results

have your script listen for a magic packet with your guests mac address.

the web ui is available on both networks, but i believe you have it right. the network all the nodes used to join the cluster.

is it possible to change the default network without reinstalling? by default i mean the network used for inter-node traffic such as live migration. id like to erase visibility of proxmox from the main network and use the faster storage network for management, live migrations etc. we have three...

skeats, yes, you can have HA but there will be points of failure. the NFS/qdisk server (unless thats already an HA setup), and the network switch are two of them. switches rarely fail. if either of those fail, data corruption is unlikely, since the data is now unavailable. there might be other...

third machine for quorum, not fencing. you can do fencing with idrac/ipmi etc, or with switch mangement. if you can afford it, networked power switches are the best way.

as usual proxmox has the functionality, but doesnt advertise :) man fence_pve.

as described in the troubleshooting, a serial console can be added with args: -serial unix:/var/run/qemu-server/<VMID>.serial,server,nowait why not make that default? the vga console would still exist unless the user disabled that within the vm. then qm could have a console command for linux...

I have installed ubuntu on an external hard drive from linux using qemu directly and then booting my macbook (2013 model) on it. something like this should work if the external shows up as sdb. might be able to do something similar from os x with virtualbox. kvm -m 1024 -drive...

were about to go into production, and i got the clearance for the supported version. our nodes have the pve-no-subscription repo. how should we switch to supported? is a reinstall nessescary? can we do this one node at a time?

as i understand it, the point of pmxcfs is that it will tell the previously downed node that those instances are already on other nodes. if it cant see the rest of the cluster, then it would not have quorum. but if this is not the case, maybe we should take that action="on" out of the fence action.

pmxcfs is brilliant. if i ever meet whoever came up with that, im getting you beer (or booch if you dont drink) anyway, seems like it prevents a rebooted node from trying restart a migrated instance. between that, and the fact that we only have 3 nodes, so if one goes down, it should come up...

is it possible to make instances that are, by definition, not important? if the cluster needs the resources, for example during an failover migration, it can suspend or kill them? maybe send an alert, and/or kill bring them back up when the resources are available again? i think this is better...

got fencing and HA migration working, but theres some tricky parts. id like to make a playbook(ansible) to check for these. i noticed a couple things to check for. in one case rgmanager was not running for some reason. restart from gui. this prevented HA migration to that node. the error says...

Is there an online expandable filesystem for virtual storage? would like to be able to resize a guests block device on the ceph cluster and its filesystem online without downtime. is this possible?

you could put them on a shared glusterfs volume. that way, if they get more hardware for their cluster, you could simply add to it. if you want to do HA, all you need is a third dummy box to make quorum easy. this third box only needs to sit on the network and be reasonably reliable. theres...

Thats what were doing, but using an existing ceph cluster, so dont know about proxmoxes built in ceph management. fast and reliable ssds are also really expensive, so we went with journals on a small partition on each osd disk. probably not as fast, but has been fast enough for us. each ceph...

Was looking through the table of contents in Mr. Chengs new book on HA, and saw a lot of discussion of drbd. i thought drbd would be dead by now. we have ceph and glusterfs. it is apparently hard to get it right. lots of moving parts, lots of potential for mistakes. it does not scale out. if...

title says it all. i realize there are work arounds like moving storage or making a an ubuntu rbd mounting appliance. but, would be nice to just have this.

there are many api calls that require a vmid even when creating a new guest. can this be done automatically? also would be nice to refer to guests by name instead of vmid in api calls.

Thanks. proxmox is not at a state where we can run a supported version yet. i hope this changes by 3.5