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

Better handle subthread timeouts in keepalive 

Prevent the main keepalive thread from getting stuck due to a subthread
taking an enormous time. If this happens, the rest of the main keepalive
will continue onward, thus ensuring that the main keepalive does not
fail for a significant number of cycles, which would cause a fence.
This commit is contained in:
Joshua Boniface 2020-08-11 11:37:26 -04:00
parent 5526e13da9
commit 46ffe352e3
1 changed files with 20 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
node-daemon/pvcnoded/Daemon.py
View File

@ -1463,24 +1463,36 @@ def node_keepalive():
# Join against running threads # Join against running threads
if enable_hypervisor: if enable_hypervisor:
vm_stats_thread.join() vm_stats_thread.join(timeout=4.0)
if vm_stats_thread.is_alive():
logger.out('VM stats gathering exceeded 4s timeout, continuing', state='w')
if enable_storage: if enable_storage:
ceph_stats_thread.join() ceph_stats_thread.join(timeout=4.0)
if ceph_stats_thread.is_alive():
logger.out('Ceph stats gathering exceeded 4s timeout, continuing', state='w')
# Get information from thread queues # Get information from thread queues
if enable_hypervisor: if enable_hypervisor:
try:
this_node.domains_count = vm_thread_queue.get() this_node.domains_count = vm_thread_queue.get()
this_node.memalloc = vm_thread_queue.get() this_node.memalloc = vm_thread_queue.get()
this_node.vcpualloc = vm_thread_queue.get() this_node.vcpualloc = vm_thread_queue.get()
except:
pass
else: else:
this_node.domains_count = 0 this_node.domains_count = 0
this_node.memalloc = 0 this_node.memalloc = 0
this_node.vcpualloc = 0 this_node.vcpualloc = 0
if enable_storage: if enable_storage:
try:
ceph_health_colour = ceph_thread_queue.get() ceph_health_colour = ceph_thread_queue.get()
ceph_health = ceph_thread_queue.get() ceph_health = ceph_thread_queue.get()
osds_this_node = ceph_thread_queue.get() osds_this_node = ceph_thread_queue.get()
except:
ceph_health_colour = fmt_cyan
ceph_health = 'UNKNOWN'
osds_this_node = '?'
# Set our information in zookeeper # Set our information in zookeeper
keepalive_time = int(time.time()) keepalive_time = int(time.time())