How to increase oned.log per line print length?

Community Support General
1

HI, Quick question I hope.

How can we increase the oned.log file per line printing to not truncated the output?

Example:


Fri Jan 17 13:09:42 2025 [Z0][ReM][D]: Req:8912 UID:0 IP:10.225.100.63 one.vm.attach invoked , 497, "disk=/var/lib/libvir..."

Fri Jan 17 13:10:21 2025 [Z0][ReM][D]: Req:1760 UID:0 one.zone.raftstatus result SUCCESS, "<RAFT><SERVER_ID>-1<..."

On different service fireedge there is thje below configurable for this. On oned.config I don’t seen anything similar. As a long shot I tried adding “truncate_max_length:” to oned.config, but it failed to start (as likely expected on unknown configurable.)


oneadmin@:/var/log/one$ less /etc/one/fireedge-server.conf 


# Log debug level: https://github.com/winstonjs/winston#using-logging-levels
#   0 = ERROR, 1 = WARNING, 2 = INFO, 5 = DEBUG
#
debug_level: 5

# Maximum length of log messages (chars)
# Messages exceeding this limit will be truncated
# -1 => No limit
truncate_max_length: 1000
2

Hi,

Check the parameters of the LOG_CALL_FORMAT variable in the /etc/one/oned.conf.

Usually, you’d need to uncomment the line and alter the last (%l) argument.

Best Regards,
Anton Todorov

1 Like
3

Hi Anton,

Thanks for the quick replay. I had missed that section of the oned.conf file.

I made this change and restarted the service, but given the text “invoke” on the format string; I wonder if this will also affect the response log printing? Is there another setting for responses?

Before:

#  LOG_CALL_FORMAT: Format string to log XML-RPC calls. Interpreted strings:
#     %i -- request id
#     %m -- method name
#     %u -- user id
#     %U -- user name
#     %l[number] -- param list and number of characters (optional) to print
#                   each parameter, default is 20. Example: %l300
#     %p -- user password
#     %g -- group id
#     %G -- group name
#     %a -- auth token
#     %A -- client IP address (only IPv4 supported)
#     %P -- client TCP port
#     %% -- %
#*******************************************************************************

#MAX_CONN           = 15
#MAX_CONN_BACKLOG   = 15
#KEEPALIVE_TIMEOUT  = 15
#KEEPALIVE_MAX_CONN = 30
#TIMEOUT            = 15
#RPC_LOG            = NO
#MESSAGE_SIZE       = 1073741824
#LOG_CALL_FORMAT    = "Req:%i UID:%u IP:%A %m invoked %l20"

Changed only this 1 line:

LOG_CALL_FORMAT    = "Req:%i UID:%u IP:%A %m invoked %l300"

systemctl restart opennebula.service

still “result” log is truncated:


Tue Jan 21 14:31:33 2025 [Z0][ReM][D]: Req:4288 UID:0 IP:127.0.0.1 one.vmpool.infoextended invoked , -2, -1, -1, -1
Tue Jan 21 14:31:33 2025 [Z0][ReM][D]: Req:4288 UID:0 one.vmpool.infoextended result SUCCESS, "<VM_POOL><VM><ID>508...

Also I’m currently using version 6.10.0 and attempted this config change on two different setups. 1 setup built with minione ansible playbooks and a 2nd build with OneDeploy playbooks. Which I assume covers this documentation warning?

"This functionality is only available when compiled with xmlrpc-c libraries >= 1.32. Currently only the packages distributed by OpenNebula are linked with this library."

4

Digging in the code, only the log of the request takes into account the %l parameter. The response is working with the hard-coded value of 20.

If you need the request/response to debug/develop an application using the OpenNebula API I found it useful to set mitmproxy and observe the full interaction from there.

Best Regards,
Anton Todorov