When creating the nbd, we wait until it's ready, but we didn't do
that when loading the configuration from JSON, which resulted in
sporadic IO failures, as the device hasn't been initialized yet. This
patch adds waitfornbd after each load_config call.
Change-Id: Id350ae7b1afab11f5f3fbd131d938dbd65a8cb15
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/459616
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Both clean and dirty shutdown tests were expanded to perform additional
writes after restore, to make sure write pointers have been correctly
re-initialised and don't cause data corruption.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: I662cf196319a39fb1cab455f5a76571904c20215
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457618
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Add test which causes dirty shutdown of ftl and attempts to restore
it. Without persistent cache it expects partial recovery of data
and zeroed out blocks in found open bands.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: I316cbbb47d0d03b2c20247df43ba789f2fd018db
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456470
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
