| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add support for tiptorrent image transfer mode to client's bash scripts
(aka cloning engine). If desired, clients receive images to restore with
the new unicast transfer mode tiptorrent.
See also: http://git.soleta.eu/tiptorrent/
#1062 Add tiptorrent support to WebConsole
Add tiptorrent option to WebConsole restore image form.
#1062 Launch tiptorrent on oglive init
This commit adds tiptorrent launch to oglive init process. Clients
launch tiptorrent daemon if they have cache partition.
#1062 Fix minor tiptorrent integration bugs
* Fix md5sum pipe typo
* Use correct variable for checksum download
#1062 Calculate checksum on image creation
Otherwise, OpenGnsys depends on cron and bittorrent specific scripts to
obtain images checksums.
Support cache cleaning for tiptorrent image parts
Tiptorrent downloads and saves images to the cache in parts. With this
commit, WebConsole shows and deletes all parts as one.
#1062 Rename TIPTORRENT-CACHE in restore image form
This commit changes TIPTORRENT for TIPTORRENT-CACHE to keep naming
consistency.
#1062 tiptorrent stores whole file instead of chunks
Tiptorrent now stores the whole file instead of the split chunks.
Removes any chunk related code from cloning engine scripts.
Requires newer version of tiptorrent installed
(tiptorrent-static >= 1.0.0-6).
Always compute MD5 of cached image
Do not rely on MD5 checksum file from the cache on tiptorrent transfers.
Otherwise, cloning engine could restore a corrupted image.
#1062 Check tiptorrent exit code
Otherwise, the image restoration continues in failed downloads.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New python requirements of programs running in ogLive clients
are shared via samba. Appends this Samba directory to the PYTHONPATH of
OpenGnsys clients (ogLive).
Specific python versions folder in the OpenGnsys server are located at:
/opt/opengnsys/client/lib/pythonX.Y/
X = Major version
Y = Minor version
This is exposed in the client (ogLive) at:
/opt/opengnsys/lib/pythonX.Y/
Clients startup script detects python3 version and exports the correct
samba directory.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before starting ogClient, script will lookup for specific classroom
configuration file, if it is found, then it will use it, otherwise it
uses the default configuration file.
The specific configuration file allows for finer-grain configurations
per room. Spaces in the room name are replaced by underscore.
For example. client: pc11, room name: "Virtual room".
If "Virtual_room.json" exists in client /opt/opengnsys/ogClient/cfg
it will use that config file.
If it doesn't exist, start ogclient with default config file
(ogclient.json)
|
| |
|
|
|
|
| |
boot of NVRAM.
|
|
|
|
| |
Windows (UEFI).
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
poweroffconf is checking if the system is idle by doing a pgrep of
processes starting with "/opt/opengnsys/" and ignoring the ones
containing $OGETC or "poweroffconf".
ogclient process always falls under this process lookup thus making
the system seem as always active.
Ignore the ogclient process when checking if the system is not idle.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's been reported that some UEFI implementations of several
manufacturers do not comply with the standard and behave oddly.
In this case it has been noted that when a BootEntry was created
(with efibootmgr) but not added to the BootOrder this new entry did not
survive a reboot. Thus, UEFI booting of a client from the ogLive menu was not
possible.
Adjust calling ogNvramAddEntry so that it adds the new entry to the
BootOrder, but in last place so network boot, usually first option,
remains in place.
|
| |
|
|
|
|
| |
in the boot menu.
|
|
|
|
|
|
|
| |
Extend database to store new fields that contain information about "clonator",
"compressor", "filesystem" and "datasize".
This patch also creates a /tmp/image.info file that is consumed by ogClient.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Irina reports that booting an OS from ogLive is not working. This is happening
because the bootOS shell script kills ogClient before invoking kexec to start
the OS. However, ogClient runs this shell script that initiates the boot
process, so it is also stopping the boot process.
The ogClient needs to close the connection with ogServer before kexec,
otherwise ogServer keeps reporting via REST API that the client is in ogLive
state / busy state. Because after kexec everything is gone basically.
This patch updates the script to:
1) spawn a process to kill the ogClient in 2 seconds, so the ogClient has time
to reply to the ogServer and to close the connection gracefully.
2) spawn a process to invoke kexec in 3 seconds to boot the OS.
[1] 2ead1639fd54bcf92d56cf7bf89b8d9cf4d6d24e
|
|
|
|
| |
It has been replaced by the ogClient.
|
|
|
|
|
| |
When the ogLive boot an OS with kexec, it did not kill gracefully
ogClient. This provokes a hanging connection in the ogServer.
|
|
|
|
| |
Instead of invoking python3.
|
|
|
|
| |
This has been renamed from ogClient/main.py to ogClient/ogclient.
|
|
|
|
| |
Step forward to rename all ogAdmServer references to ogServer
|
|\ |
|
| | |
|
|/
|
|
|
|
|
| |
This commit changes the bash script that the clients execute at the init
fase to launch the new ogClient instead of the old ogAdmClient. This way
the client works with the new ogAdmServer that do not use custom
protocol SocketHidra anymore.
|
|
|
| |
OGAgent code, browser code and the ogLive builder scrips are moved to new Git repositories.
|
|
|
| |
Fix bug introduced by commit `930563c`due to the change of the database encoding.
|
|
|
|
| |
function is created for detect MBR content and used in wizard.
|
|
|
|
| |
it is created.
|
|
|
|
| |
of OGAgent for Windows.
|
| |
|
| |
|
|
|
|
| |
condition to match the whole line.
|
|
|
|
|
| |
ogGrub4dosInstallMbr: the configuration directory name is changed to lowercase for the script to work in Linux.
GRUBDIR is removed because not use.
|
| |
|
| |
|
|
|
|
| |
entries or access to command line.
|
| |
|
|
|
|
| |
MSDOS disk MBR
|
|
|
|
|
|
|
|
|
| |
firmware
OG_ERR_NOTBIOS
error code: loadenviron.sh
string code: lang_es:cat:gb
admin code: ogRaiseError () fron system.lib
|
|
|
|
|
|
|
|
|
| |
MSDOS
OG_ERR_NOMSDOS
error code: loadenviron.sh
string code: lang_es:cat:gb
admin code: ogRaiseError () fron system.lib
|
| |
|
|
|
|
|
| |
Check the availability of the binary in the ogLive. If necessary use the sources from $ OGLIB / burg / burg.tgz
Live. If necessary use the sources from $ OGLIB / burg / burg.tgz
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
created with other partclone version.
|
|
|
|
| |
no es la adecuada para arrancar este tipo de discos, se detecta y renombra la entrada del disco correspondiente con la nomenclatura: hd0,0 para discos nvme con tabla MSDOS y hd0,gpt1 para discos nvme con tabla GPT
|
|
|
|
| |
better detected, avoiding errors.
|
|
|
|
| |
drive. You can use the rEFInd template example for computers without SecureBoot.
|