]>
Commit | Line | Data |
---|---|---|
1 | ifdef::manvolnum[] | |
2 | pmxcfs(8) | |
3 | ========= | |
4 | include::attributes.txt[] | |
5 | :pve-toplevel: | |
6 | ||
7 | NAME | |
8 | ---- | |
9 | ||
10 | pmxcfs - Proxmox Cluster File System | |
11 | ||
12 | SYNOPSIS | |
13 | -------- | |
14 | ||
15 | include::pmxcfs.8-cli.adoc[] | |
16 | ||
17 | DESCRIPTION | |
18 | ----------- | |
19 | endif::manvolnum[] | |
20 | ||
21 | ifndef::manvolnum[] | |
22 | Proxmox Cluster File System (pmxcfs) | |
23 | ==================================== | |
24 | include::attributes.txt[] | |
25 | endif::manvolnum[] | |
26 | ifdef::wiki[] | |
27 | :pve-toplevel: | |
28 | endif::wiki[] | |
29 | ||
30 | The Proxmox Cluster file system (``pmxcfs'') is a database-driven file | |
31 | system for storing configuration files, replicated in real time to all | |
32 | cluster nodes using `corosync`. We use this to store all PVE related | |
33 | configuration files. | |
34 | ||
35 | Although the file system stores all data inside a persistent database | |
36 | on disk, a copy of the data resides in RAM. That imposes restriction | |
37 | on the maximum size, which is currently 30MB. This is still enough to | |
38 | store the configuration of several thousand virtual machines. | |
39 | ||
40 | This system provides the following advantages: | |
41 | ||
42 | * seamless replication of all configuration to all nodes in real time | |
43 | * provides strong consistency checks to avoid duplicate VM IDs | |
44 | * read-only when a node loses quorum | |
45 | * automatic updates of the corosync cluster configuration to all nodes | |
46 | * includes a distributed locking mechanism | |
47 | ||
48 | ||
49 | POSIX Compatibility | |
50 | ------------------- | |
51 | ||
52 | The file system is based on FUSE, so the behavior is POSIX like. But | |
53 | some feature are simply not implemented, because we do not need them: | |
54 | ||
55 | * you can just generate normal files and directories, but no symbolic | |
56 | links, ... | |
57 | ||
58 | * you can't rename non-empty directories (because this makes it easier | |
59 | to guarantee that VMIDs are unique). | |
60 | ||
61 | * you can't change file permissions (permissions are based on path) | |
62 | ||
63 | * `O_EXCL` creates were not atomic (like old NFS) | |
64 | ||
65 | * `O_TRUNC` creates are not atomic (FUSE restriction) | |
66 | ||
67 | ||
68 | File Access Rights | |
69 | ------------------ | |
70 | ||
71 | All files and directories are owned by user `root` and have group | |
72 | `www-data`. Only root has write permissions, but group `www-data` can | |
73 | read most files. Files below the following paths: | |
74 | ||
75 | /etc/pve/priv/ | |
76 | /etc/pve/nodes/${NAME}/priv/ | |
77 | ||
78 | are only accessible by root. | |
79 | ||
80 | ||
81 | Technology | |
82 | ---------- | |
83 | ||
84 | We use the http://www.corosync.org[Corosync Cluster Engine] for | |
85 | cluster communication, and http://www.sqlite.org[SQlite] for the | |
86 | database file. The file system is implemented in user space using | |
87 | http://fuse.sourceforge.net[FUSE]. | |
88 | ||
89 | File System Layout | |
90 | ------------------ | |
91 | ||
92 | The file system is mounted at: | |
93 | ||
94 | /etc/pve | |
95 | ||
96 | Files | |
97 | ~~~~~ | |
98 | ||
99 | [width="100%",cols="m,d"] | |
100 | |======= | |
101 | |`corosync.conf` | Corosync cluster configuration file (previous to {pve} 4.x this file was called cluster.conf) | |
102 | |`storage.cfg` | {pve} storage configuration | |
103 | |`datacenter.cfg` | {pve} datacenter wide configuration (keyboard layout, proxy, ...) | |
104 | |`user.cfg` | {pve} access control configuration (users/groups/...) | |
105 | |`domains.cfg` | {pve} authentication domains | |
106 | |`authkey.pub` | Public key used by ticket system | |
107 | |`pve-root-ca.pem` | Public certificate of cluster CA | |
108 | |`priv/shadow.cfg` | Shadow password file | |
109 | |`priv/authkey.key` | Private key used by ticket system | |
110 | |`priv/pve-root-ca.key` | Private key of cluster CA | |
111 | |`nodes/<NAME>/pve-ssl.pem` | Public SSL certificate for web server (signed by cluster CA) | |
112 | |`nodes/<NAME>/pve-ssl.key` | Private SSL key for `pve-ssl.pem` | |
113 | |`nodes/<NAME>/pveproxy-ssl.pem` | Public SSL certificate (chain) for web server (optional override for `pve-ssl.pem`) | |
114 | |`nodes/<NAME>/pveproxy-ssl.key` | Private SSL key for `pveproxy-ssl.pem` (optional) | |
115 | |`nodes/<NAME>/qemu-server/<VMID>.conf` | VM configuration data for KVM VMs | |
116 | |`nodes/<NAME>/lxc/<VMID>.conf` | VM configuration data for LXC containers | |
117 | |`firewall/cluster.fw` | Firewall configuration applied to all nodes | |
118 | |`firewall/<NAME>.fw` | Firewall configuration for individual nodes | |
119 | |`firewall/<VMID>.fw` | Firewall configuration for VMs and Containers | |
120 | |======= | |
121 | ||
122 | ||
123 | Symbolic links | |
124 | ~~~~~~~~~~~~~~ | |
125 | ||
126 | [width="100%",cols="m,m"] | |
127 | |======= | |
128 | |`local` | `nodes/<LOCAL_HOST_NAME>` | |
129 | |`qemu-server` | `nodes/<LOCAL_HOST_NAME>/qemu-server/` | |
130 | |`lxc` | `nodes/<LOCAL_HOST_NAME>/lxc/` | |
131 | |======= | |
132 | ||
133 | ||
134 | Special status files for debugging (JSON) | |
135 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
136 | ||
137 | [width="100%",cols="m,d"] | |
138 | |======= | |
139 | |`.version` |File versions (to detect file modifications) | |
140 | |`.members` |Info about cluster members | |
141 | |`.vmlist` |List of all VMs | |
142 | |`.clusterlog` |Cluster log (last 50 entries) | |
143 | |`.rrd` |RRD data (most recent entries) | |
144 | |======= | |
145 | ||
146 | ||
147 | Enable/Disable debugging | |
148 | ~~~~~~~~~~~~~~~~~~~~~~~~ | |
149 | ||
150 | You can enable verbose syslog messages with: | |
151 | ||
152 | echo "1" >/etc/pve/.debug | |
153 | ||
154 | And disable verbose syslog messages with: | |
155 | ||
156 | echo "0" >/etc/pve/.debug | |
157 | ||
158 | ||
159 | Recovery | |
160 | -------- | |
161 | ||
162 | If you have major problems with your Proxmox VE host, e.g. hardware | |
163 | issues, it could be helpful to just copy the pmxcfs database file | |
164 | `/var/lib/pve-cluster/config.db` and move it to a new Proxmox VE | |
165 | host. On the new host (with nothing running), you need to stop the | |
166 | `pve-cluster` service and replace the `config.db` file (needed permissions | |
167 | `0600`). Second, adapt `/etc/hostname` and `/etc/hosts` according to the | |
168 | lost Proxmox VE host, then reboot and check. (And don't forget your | |
169 | VM/CT data) | |
170 | ||
171 | ||
172 | Remove Cluster configuration | |
173 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
174 | ||
175 | The recommended way is to reinstall the node after you removed it from | |
176 | your cluster. This makes sure that all secret cluster/ssh keys and any | |
177 | shared configuration data is destroyed. | |
178 | ||
179 | In some cases, you might prefer to put a node back to local mode | |
180 | without reinstall, which is described here: | |
181 | ||
182 | * stop the cluster file system in `/etc/pve/` | |
183 | ||
184 | # systemctl stop pve-cluster | |
185 | ||
186 | * start it again but forcing local mode | |
187 | ||
188 | # pmxcfs -l | |
189 | ||
190 | * remove the cluster configuration | |
191 | ||
192 | # rm /etc/pve/cluster.conf | |
193 | # rm /etc/cluster/cluster.conf | |
194 | # rm /var/lib/pve-cluster/corosync.authkey | |
195 | ||
196 | * stop the cluster file system again | |
197 | ||
198 | # systemctl stop pve-cluster | |
199 | ||
200 | * restart PVE services (or reboot) | |
201 | ||
202 | # systemctl start pve-cluster | |
203 | # systemctl restart pvedaemon | |
204 | # systemctl restart pveproxy | |
205 | # systemctl restart pvestatd | |
206 | ||
207 | ||
208 | ifdef::manvolnum[] | |
209 | include::pve-copyright.adoc[] | |
210 | endif::manvolnum[] |