You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Today rflash command uploads files in parallel by forking a process for each node. If upload operation fails the error is saved in node_info hash. At the end the summary is displayed for the attempted range of nodes and errors are displayed for each failed node.
However, since each upload happens in a separate process, the node_info is just a copy, so that when the upload finished, the node_info hash does not have the errors saved by each process. And we display generic "BMC is not ready" message.
The solution is probably to return the error string from forked process instead of 1 if there is an upload error. Hopefully the returned string can be caught by the main process and saved into node_info hash.
[root@briggs01 gurevich]# rflash mid05tor12cn15 -a /tmp/gurevich/fw.tar
Attempting to upload /tmp/gurevich/fw.tar, please wait...
mid05tor12cn15: Failed to upload update file /tmp/gurevich/fw.tar : 400 Bad Request - Version already exists or failed to be extracted
-------------------------------------------------------
Firmware update complete: Total=1 Success=0 Failed=1
mid05tor12cn15: BMC is not ready
-------------------------------------------------------
[root@briggs01 gurevich]#
The text was updated successfully, but these errors were encountered:
Today
rflash
command uploads files in parallel by forking a process for each node. If upload operation fails the error is saved innode_info
hash. At the end the summary is displayed for the attempted range of nodes and errors are displayed for each failed node.However, since each upload happens in a separate process, the
node_info
is just a copy, so that when the upload finished, thenode_info
hash does not have the errors saved by each process. And we display generic "BMC is not ready" message.The solution is probably to return the error string from forked process instead of
1
if there is an upload error. Hopefully the returned string can be caught by the main process and saved intonode_info
hash.The text was updated successfully, but these errors were encountered: