parallelvirtualcluster
/
pvc
1
0
Fork 0
Code Issues 11 Pull Requests Projects Releases 111 Wiki Activity

Refactor OSD removal to use new ZK data 

With the OSD LVM information stored in Zookeeper, we can use this to
determine the actual block device to zap rather than relying on runtime
determination and guestimation.
This commit is contained in:
Joshua Boniface 2022-05-02 12:52:22 -04:00
parent 464f0e0356
commit c741900baf
1 changed files with 36 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

71
node-daemon/pvcnoded/objects/CephInstance.py
View File

@ -542,47 +542,48 @@ class CephOSDInstance(object):
break break
# 4. Determine the block devices # 4. Determine the block devices
device_zk = zkhandler.read(("osd.device", osd_id)) osd_vg = zkhandler.read(("osd.vg", osd_id))
try: osd_lv = zkhandler.read(("osd.lv", osd_id))
retcode, stdout, stderr = common.run_os_command( osd_lvm = f"/dev/{osd_vg}/{osd_lv}"
"readlink /var/lib/ceph/osd/ceph-{}/block".format(osd_id) osd_device = None
)
vg_name = stdout.split("/")[
-2
] # e.g. /dev/ceph-<uuid>/osd-block-<uuid>
retcode, stdout, stderr = common.run_os_command(
"vgs --separator , --noheadings -o pv_name {}".format(vg_name)
)
pv_block = stdout.strip()
except Exception as e:
print(e)
pv_block = device_zk
# 5a. Verify that the blockdev actually has a ceph volume that matches the ID, otherwise don't zap it
logger.out( logger.out(
f"Check OSD disk {pv_block} for OSD signature with ID osd.{osd_id}", f"Getting disk info for OSD {osd_id} LV {osd_lvm}",
state="i", state="i",
) )
retcode, stdout, stderr = common.run_os_command( retcode, stdout, stderr = common.run_os_command(
f"ceph-volume lvm list {pv_block}" f"ceph-volume lvm list {osd_lvm}"
) )
if f"====== osd.{osd_id} =======" in stdout: for line in stdout.split("\n"):
# 5b. Zap the volumes if "devices" in line:
logger.out( osd_device = line.split()[-1]
"Zapping OSD disk with ID {} on {}".format(osd_id, pv_block),
state="i", if not osd_device:
) print("ceph-volume lvm list")
retcode, stdout, stderr = common.run_os_command( print("Could not find OSD information in data:")
"ceph-volume lvm zap --destroy {}".format(pv_block) print(stdout)
) print(stderr)
if retcode: if force_flag:
print("ceph-volume lvm zap") logger.out("Ignoring error due to force flag", state="i")
print(stdout) else:
print(stderr) raise Exception
if force_flag:
logger.out("Ignoring error due to force flag", state="i") # 5. Zap the volumes
else: logger.out(
raise Exception "Zapping OSD {} disk on {}".format(osd_id, osd_device),
state="i",
)
retcode, stdout, stderr = common.run_os_command(
"ceph-volume lvm zap --destroy {}".format(osd_device)
)
if retcode:
print("ceph-volume lvm zap")
print(stdout)
print(stderr)
if force_flag:
logger.out("Ignoring error due to force flag", state="i")
else:
raise Exception
# 6. Purge the OSD from Ceph # 6. Purge the OSD from Ceph
logger.out("Purging OSD disk with ID {}".format(osd_id), state="i") logger.out("Purging OSD disk with ID {}".format(osd_id), state="i")