Scope
The 1.3.x release of MPX Node had a small bug where log files did not rotate. This caused storage space on the device to fill to 100% and could cause the web interface to crash, however MPX will still encode/decode. This has been fixed in v1.5 but the problem will reman if the log space wasn't cleared before the update. Most users never run into this issue as logs never grow large enough. In some cases, the unit may run into certain error conditions and the log size will grow quickly
Do not share this document with customers - this document is for support reference only
Steps to clear out log files
1) Start a remote session with the customer's PC that can access the MPX Nodes
2) Using PuTTY or other SSH client, connect to MPX node via SSH (root, "Omnia4.5" in v1.3.x or earlier) (root, "xeazv9gQjcfjHwZ54ffD" in v1.5 or later)
3) Run command df -h
Filesystem Size Used Available Use% Mounted on
devtmpfs 453.5M 0 453.5M 0% /dev
tmpfs 497.8M 0 497.8M 0% /dev/shm
tmpfs 497.8M 40.0K 497.7M 0% /tmp
tmpfs 497.8M 32.0K 497.8M 0% /run
/dev/mmcblk2p1 118.2M 84.5M 27.6M 75% /bank0
/dev/mmcblk2p2 118.2M 84.5M 27.6M 75% /bank1
/dev/mmcblk2p3 938.7M 11.6M 879.5M 1% /config
Observe /config in the "Mounted on" column. If its space is completely used or its approaching 100% its clear that unit needs space reclaimed
4) Switch into the /config directory cd /config/MPXnode/uMPX/
5) The logs directory is hidden. Use ls -al
to show hidden files
cd .MicroMPX_Decoder.log
If this is a Decoder unit logs are here
cd .MicroMPX_Encoder.log
If this is a Encoder unit logs are here
6) Once you are in the Encoder or Decoder log directory run rm *.log*
to delete all log files
7) Re-run df -h command to confirm space has been reclaimed
**If the web server was already crashed when you logged into the unit, after freeing the space you will need to reboot the MPX node. **
If /config still reports 100% used space
If this did not work and the /config directly still shows 100% use, the file system is likely corrupted. Contact Ryan W. in engineering for help.
Below are the steps that would be advised by engineering to fix filesystem errors:
1) Remain logged into MPX node as root user
2) Enter this command into terminal,killall restart.sh
, press enter to execute
2) Copy and Paste the following into the terminal and press enter to execute
PROCS=$(lsof |grep /config | grep -v grep | awk ' { printf " "$1 } ')
echo "PROCS: $PROCS"
for PID in "$PROCS"
do
kill -15 "$PID"
done
3)Copy and Paste the following into the terminal and press enter to execute
PROCS=$(ps auxw|grep mppd | grep -v grep | awk ' { printf " "$2 } ')
echo "PROCS: $PROCS"
for PID in "$PROCS"
do
kill -15 "$PID"
done
4) Enter this command into terminal, umount /config
and press enter to execute
5) STOP. If Step 4 did not report any errors only continue on to this step. If there were errors, stop here and contact engineering
Enter this command into terminal, fsck -y /config
and press enter to execute
6) Enter this command into terminal, find /config/MPXnode -type f -size 0 -exec rm {} \;
and press enter to execute. (finds any zero byte files and deletes)
7) If all above commands executed without errors reboot unit with reboot
command and confirm free space is availabe again for /config mount