Verify that yast network module doesn't crash in case of wrong configuration in ifcfg
This test scenario occurred in the customer setup, see https://bugzilla.suse.com/show_bug.cgi?id=1061306
YaST must not break if it finds something unexpected in network configuration files (ifcfg). The fact is, that if anything changes in the future, we need to be ready. Customers can add some values either by a mistake or even intentionally, for instance, if they want/need it for some reason (that might be unknown to us).
Scope is 64bit only. We can add this scenario to yast2_gui test suite.
Test scenario should include at least following steps:
- Put misconfigured settings in ifcfg file e.g. ETHERDEVICE setting which is expected to be used for VLANs only: ETHERDEVICE='eth0'
- Run lan module of the yast nd verify it doesn't crash
- Implement automated test
- Add to the SLE Functional job group
- Automated test runs in production and verifies that lan yast module doesn't crash when ifcfg file contains invalid configuration
#3 Updated by okurz almost 3 years ago
- Subject changed from [sle][functional][yast] Create test to verify that yast network module doesn't crash in case of wrong configuration in ifcfg to [sle][functional][yast][y] Create test to verify that yast network module doesn't crash in case of wrong configuration in ifcfg
- Target version changed from Milestone 16 to Milestone 18
M16 seems to be too ambitious
- Project changed from openQA Tests to qe-yast
- Subject changed from [sle][functional][yast][y] Create test to verify that yast network module doesn't crash in case of wrong configuration in ifcfg to Verify that yast network module doesn't crash in case of wrong configuration in ifcfg
- Category deleted (
#9 Updated by JRivrain about 1 month ago
Nowadays, ETHERDEVICE='eth0' in ifcfg-eth0 marks eth0 as parent device of eth0, or whatever we wrote between quotes (eg ETHERDEVICE='Hahahasomuchfun'). Does not crash, but then we have a different problem, probably not a bug though. If the purpose is to see whether putting a wrong entry provokes a crash, we should find something else probably.
Otherwise if we put a typo like "THERDEVICE='eth0'", the entry gets ignored. I guess testing this may be relevant in the scope of this ticket. Scope that otherwise I'm having hard time to define here, as there can be litteraly an inifinity of mistakes in a hand-written config file.
I guess a good starting point can be that script from the bug report https://bugzilla.suse.com/attachment.cgi?id=746128