[pve-devel] trying to upgrade wheezy -> jessie + corosync1 ->corosync2

Alexandre DERUMIER aderumier at odiso.com
Tue Apr 28 17:31:13 CEST 2015


>>a)I can see other nodes green. 
>>
>>Seem that pvestatd cannot send through corosync status to other nodes. 
>>
>>
>>b) editing a file in /etc/pve/ in one node, doesn't apply on other nodes 

Fixed by upgrade pve-cluster on both node to last version compiled from git



root at kvmtest1 ~/proxmox2/pve-cluster (master) # dpkg -i pve-cluster_4.0-8_amd64.deb 
(Reading database ... 363622 files and directories currently installed.)
Preparing to unpack pve-cluster_4.0-8_amd64.deb ...
Unpacking pve-cluster (4.0-8) over (4.0-8) ...
Setting up pve-cluster (4.0-8) ...
insserv: script suricata2: service suricata already provided!
Restarting pve cluster filesystem: pve-cluster.
Processing triggers for systemd (215-17) ...
Processing triggers for pve-firewall (1.0-17) ...
Restarting Proxmox VE firewall: pve-firewall.
Processing triggers for man-db (2.7.0.2-5) ...



Now replication in /etc/pve works fine, and I'm seeing all nodes green :)

(All without need to reboot wheezy or restart vm :)


Ready for testing live migration.



----- Mail original -----
De: "aderumier" <aderumier at odiso.com>
À: "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Mardi 28 Avril 2015 16:55:48
Objet: Re: [pve-devel] trying to upgrade wheezy -> jessie + corosync1	->corosync2

>>c) try to use #qm stop|start vmid 
>> 
>>give me "unknown file type 'cluster.conf'" 

This one come from old version of qemu-server in proxmox jessie repo 

my $vm_is_ha_managed = sub { 
my ($vmid) = @_; 

my $cc = PVE::Cluster::cfs_read_file('cluster.conf'); 
if (PVE::Cluster::cluster_conf_lookup_pvevm($cc, 0, $vmid, 1)) { 
return 1; 
} 
return 0; 
}; 


can you update the repo ? 


----- Mail original ----- 
De: "aderumier" <aderumier at odiso.com> 
À: "pve-devel" <pve-devel at pve.proxmox.com> 
Envoyé: Mardi 28 Avril 2015 16:40:09 
Objet: [pve-devel] trying to upgrade wheezy -> jessie + corosync1 ->corosync2 

Hi, 

I'm begin to test inplace upgrade from wheezy to jessie + corosync1->corosync2 without reboot 

here some note: 

1)-update repos to jessie 

2)apt-get dist-upgrade 

3)aptitude remove redhat-cluster-pve fence-agents-pve redhat-cluster-pve-dev resource-agents-pve clvm dlm3 

Remove the following packages: 
1) fence-agents-pve 
2) libfile-fcntllock-perl 
3) redhat-cluster-pve 
4) redhat-cluster-pve-dev 
5) resource-agents-pve 

(I need to use aptitude because of conflict with libdlm3) 

4)killall -9 corosync 

5)apt-get install libcorosync4-pve 

6)generate a /etc/corosync/authkey 

7) convert cluster.conf to /etc/corosync/corosync.conf 

totem { 
version: 2 
secauth: on 
cluster_name: clustertest 
config_version: 1 
} 

nodelist { 
node { 
ring0_addr: kvmtest1 
nodeid: 1 
quorum_votes: 1 
} 

node { 
ring0_addr: kvmtest2 
nodeid: 2 
quorum_votes: 1 
} 

node { 
ring0_addr: kvmtest3 
nodeid: 3 
quorum_votes: 1 
} 


} 

quorum { 
provider: corosync_votequorum 
} 

logging { 
to_syslog: yes 
debug: off 
} 


8) start corosync 
#corosync 


9)-stop/start all proxmox daemons (with old /etc/init.d/pve-* scripts) 



All seem to works fine, corosync seeing each nodes, 
/etc/pve is mounted 
On each node, I can see the local node green 

But.... 

a)I can see other nodes green. 

Seem that pvestatd cannot send through corosync status to other nodes. 


b) editing a file in /etc/pve/ in one node, doesn't apply on other nodes 


c) try to use #qm stop|start vmid 

give me "unknown file type 'cluster.conf'" 


Any idea how to debug that ? 



_______________________________________________ 
pve-devel mailing list 
pve-devel at pve.proxmox.com 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel 



More information about the pve-devel mailing list