Pantavisor Configuration¶
This reference is for advanced users. Before manually editing these values in your config, check if what you intend to do can be achieved through one of the build options.
There are three ways to tweak with Pantavisor configuration, depending on when it can be modified: at compile time, after an update or during runtime.
At compile time¶
You can take a look at the configuration files we use to build our supported platforms here.
pantavisor.config¶
| Key | Value | Default | Description |
|---|---|---|---|
| dropbear.cache.dir | path | /storage/cache/dropbear | set dropbear cache path |
| meta.cache.dir | path | /storage/cache/meta | set metadata cache |
| bootloader.type | uboot, uboot-pvk or grub | uboot | set type of bootloader |
| bootloader.mtd_only | 0 or 1 | 0 | enable MTD |
| bootloader.mtd_env | path | N/A | set MTD path |
| storage.device | path | N/A | set storage path |
| storage.fstype | ext4, ubifs or jffs2 | N/A (mandatory) | journaling file system |
| storage.opts | N/A | N/A | not implemented |
| storage.mntpoint | path | N/A (mandatory) | set pv mount point |
| storage.mnttype | ext4 | disabled | pv mount file system |
| storage.logtempsize | size | disabled | sets the size of memory dedicated for log storage. The size may have a k, m or g suffix or a % suffix for percentage of memory |
| storage.wait | integer | 5 | wait for storage to be available |
| storage.gc.reserved | integer | 5 | garbage collector will try to always keep this percentage of disk free |
| storage.gc.keep_factory | 0 or 1 | 0 | avoid deletion of revision 0 objects by the garbage collector |
| storage.gc.threshold | integer | 0 | garbage collector will be triggered if there is less than this percentage of disk free. 0 means it will just be run when an update needs more space than available |
| updater.use_tmp_objects | 0 or 1 | 0 | download objects in an on-disk temporary location. If disabled, objects will be stored in memory while downloading |
| revision.retries | integer | 10 | max number of update retries before rollback |
| revision.retries.timeout | integer | 120 | set time in seconds before a failed update is retired |
| wdt.enabled | 0 or 1 | 1 | enable kernel watchdog |
| wdt.timeout | integer | 15 | set reset timeout in seconds for kernel watchdog |
| net.brdev | string | lxcbr0 | container bridge name |
| net.braddress4 | ip | 10.0.3.1 | container bridge address |
| net.brmask4 | ip | 255.255.255.0 | container bridge mask address |
| lxc.log.level | integer | 2 | 0 to 8 lxc log verbosity level |
Example of pantavisor.config:
bootloader.type=normal-uboot
storage.device=LABEL=pvroot
storage.fstype=ext4
storage.opts=
storage.mntpoint=/storage
wdt.enabled=1
wdt.timeout=30
pantahub.config¶
| Key | Value | Default | Description |
|---|---|---|---|
| creds.type | builtin and ext-* | builtin | set type of authentication |
| creds.host | ip | 192.168.53.1 | Pantahub IP |
| creds.port | port | 12365 | Pantahub port |
| creds.id | string | unset (set after claim) | device id |
| creds.prn | string | unset (set after claim) | device prn |
| creds.secret | string | unset (set after claim) | device secret |
| creds.tpm.key | string | disabled | TMP key |
| creds.tpm.cert | string | disabled | TMP certificate |
| factory.autotok | string | disabled | token for auto claiming |
| updater.keep_factory | 0 or 1 | 0 | DEPRECATED: moved to storage.gc.keep_factory in pantavisor.config |
| updater.interval | integer | 60 | time between Pantahub requests during waiting operation |
| updater.network_timeout | integer | 120 | wait time in seconds before rollback if device does not get network connectivity |
| updater.commit.delay | integer | 180 | testing time in seconds after an update |
| log.dir | path | /storage/logs/ | set log storage path |
| log.maxsize | integer | 2097152 B (2 MB) | max size of a log file before rotating |
| log.level | 0 (FATAL), 1 (ERROR), 2 (WARN), 3 (INFO), 4 (DEBUG) or 5 (ALL) | 5 | pantavisor log will only capture messages from the set loglevel to 0 |
| log.buf_nitems | integer | 128 KB | log buffer size |
| log.push | 0 or 1 | 1 | push logs to cloud. Can be overriden with this metadata |
| log.capture | 0 or 1 | 1 | store logs. Disable will stop log push too |
Example of pantahub.config:
log.level=5
log.buf_nitems=128
updater.interval=5
updater.network_timeout=60
updater.commit.delay=40
creds.host=api.pantahub.com
creds.port=443
creds.id=
creds.prn=
creds.secret=
After an update¶
Some of the values from the previous section can be overridden when making a new revision of your device's firmware. To do so, you must create a new file with the same format we use for pantavisor.config and reference it from bsp/run.json.
These are the keys that are accepted in this revision configuration. See the previous section for reference: * storage.wait * storage.gc.reserved * storage.gc.keep_factory * storage.gc.threshold * updater.use_tmp_objects * revision.retries * revision.retries.timeout * updater.keep_factory * updater.interval * updater.network_timeout * updater.commit.delay * log.maxsize * log.level * log.buf_nitems * log.push * log.capture * lxc.log.level
For example, we create this file:
$ cat bsp/trail.config
log.level=3
storage.gc.threshold=30
Now, we reference it from our bsp run.json, using the key initrd_config:
$ cat bsp/run.json
{
"addons": [],
"firmware": "firmware.squashfs",
"initrd": "pantavisor",
"linux": "kernel.img",
"modules": "modules.squashfs",
"initrd_config": "trail.config"
}
To finish the process, we commit and post the new revision
$ pvr commit
Committing /home/anibal/.pvr/objects/bsp/run.json
Adding bsp/trail.config
$ pvr post -m "change loglevel and storage threshold for this revision"
During runtime¶
Metadata can be used to override the values from the previous sections.
These are the keys that are accepted in metadata configuration. See the previous section for reference: * storage.gc.reserved * storage.gc.keep_factory * storage.gc.threshold * updater.interval * log.level * log.push