Merge charm-hw-health:bug/2044387 into charm-hw-health:master

[Facundo Ciccioli]

Don't bail out when the nvme command fails

[🤖 Canonical IS Merge Bot (canonical-is-mergebot)]

This merge proposal is being monitored by mergebot. Change the status to Approved to merge.

[Robert Gildein (rgildein)]

Can you provide unit tests for these changes. Thanks
FYI: Our Jenkins is down, so CI not working now.

review: Needs Fixing

Unmerged commits

e605730... by Facundo Ciccioli on 2023-11-23

Don't bail out when the nvme command fails

Take note of it and keep going. Raise to UNKNOWN if some command failed
and no criticals were found.

Fixes bug #2044387

diff --git a/src/files/nvme/check_nvme.py b/src/files/nvme/check_nvme.py
index 14468cb..fd40071 100755
--- a/src/files/nvme/check_nvme.py
+++ b/src/files/nvme/check_nvme.py
@@ -21,6 +21,7 @@ def parse_output(warning_used, critical_used):  # noqa: C901
     keymap = {}
     critical = False
     warning = False
+    unknown = False
     alloutputs = []
     for device in glob.glob("/dev/nvme*"):
         if not NVME_RE.match(device):
@@ -31,7 +32,12 @@ def parse_output(warning_used, critical_used):  # noqa: C901
             )
         except subprocess.CalledProcessError as error:
             print("nvme check error: {}".format(error))
-            return
+            status = ("UNKNOWN: {} error running nvme command: {}").format(
+                device, error
+            )
+            unknown = True
+            alloutputs.append(status)
+            continue
 
         for line in output.decode(errors="ignore").splitlines():
             datavalues_re = re.match(r"^(\w+)\s+:\s+([\d.]+)", line.strip())
@@ -70,6 +76,9 @@ def parse_output(warning_used, critical_used):  # noqa: C901
     if critical:
         raise CriticalError("\n".join(alloutputs))
 
+    if unknown:
+        raise UnknownError("\n".join(alloutputs))
+
     if warning:
         raise WarnError("\n".join(alloutputs))