|
Behavior:
The issue is primarily encountered during the following scenarios:
- Downgrading the FortiSwitch 6XXF series from v7.4.4 (build0861) to v7.4.3 (build0830).
- Changing the firmware image using FortiSwitch flash.
diagnose sys flash list
Partition Image TotalSize(KB) Used(KB) Use% Active
1 S648FF-7.04-FW-build0830-240422-patch03 55529 43262 77% No
516040 636 0% No
(*) 2 S648FF-7.04-FW-build0861-241004-patch04 55529 41709 75% Yes
507744 2419 0% Yes
- Upon booting the FortiSwitch, an e2fsck prompt is displayed. After typing 'Y', the system shows no display for a while, followed by a crash, as demonstrated below:
EXT2-fs (sda5): warning: mounting unchecked fs, running e2fsck is recommended
Do you want to run e2fsck? Will continue to boot if no input in 2 seconds (y/n) [n] y
<01905> firmware FortiSwitch-624F-FPOE v7.04.3,build0830b830,240422 (Release)
<01905> application parse_conf
<01905> *** signal 11 (Segmentation fault) received ***
<01905> Register dump:
<01905> RAX: 0000000000000000 RBX: 0000000000000001
<01905> RCX: 00007f86a484c7a7 RDX: 0000000000000000
<01905> R8: 0000000000000000 R9: 0000000000000000
<01905> R10: 0000000000000000 R11: 0000000000000206
<01905> R12: 0000000002e81db8 R13: 0000000000000000
<01905> R14: 00007ffc52df30c8 R15: 0000000000000000
<01905> RSI: 0000000000000000 RDI: 0000000000000000
<01905> RBP: 0000000000000000 RSP: 00007ffc52dee750
<01905> RIP: 0000000000000000 EFLAGS: 0000000000010202
<01905> CSFSGS: 002b000000000033
<01905> Trap: 000000000000000e Error: 0000000000000014
<01905> OldMask: 0000000000000000
<01905> CR2: 0000000000000000
<01905> Backtrace:
- The crash can also be validated on v7.4.4 using the command provided below. The crash log specifies the build as shown below:
diagnose debug crashlog read
4: 2024-10-22 16:35:57 <01931> firmware FortiSwitch-648F-FPOE v7.04.3,build0830b830,240422 (Release)
Cause and Resolution:
The issue is caused by an error in the calculation of the QoS map size, which has been resolved in FortiSwitchOS versions 7.4.4 and 7.6.0.
Note: This issue is associated with an internally reported bug (ID: 1038763), which is not documented in the release notes.
|
