xCAT/pyghmi
2
0
Fork 0
mirror of https://opendev.org/x/pyghmi synced 2025-01-27 19:37:44 +00:00
Code Issues Projects Releases Wiki Activity

Attempt bulk request of inventory

Browse Source 
Bring more urls together in hopes of performance
through concurrency.  There are still points of
serialization, however.

Change-Id: Ie2b4eebe84cbd35e08a6fbf703d3a2b15a17026d
This commit is contained in:
Jarrod Johnson 2019-04-10 13:21:27 -04:00
parent 3475b97a91
commit 9c4e520d93
1 changed files with 43 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
pyghmi/redfish/command.py
View File

@ -550,19 +550,23 @@ class Command(object):
}
yield ('System', sysinfo)
self._hwnamemap = {}
for cpu in self._get_cpu_inventory(withids=withids):
adpurls = self._get_adp_urls()
cpurls = self._get_cpu_urls()
memurls = self._get_mem_urls()
allurls = adpurls + cpurls + memurls
list(self._do_bulk_requests(allurls))
for cpu in self._get_cpu_inventory(withids=withids, urls=cpurls):
yield cpu
for mem in self._get_mem_inventory(withids=withids):
for mem in self._get_mem_inventory(withids=withids, urls=memurls):
yield mem
for adp in self._get_adp_inventory(withids=withids):
for adp in self._get_adp_inventory(withids=withids, urls=adpurls):
yield adp
def _get_adp_inventory(self, onlyname=False, withids=False, urls=None):
if not urls:
adpurls = self.sysinfo.get('PCIeDevices', [])
if not adpurls:
return
urls = [x['@odata.id'] for x in adpurls]
urls = self._get_adp_urls()
if not urls:
return
for inf in self._do_bulk_requests(urls):
adpinfo, url = inf
aname = adpinfo.get('Name', 'Unknown')
@ -604,13 +608,19 @@ class Command(object):
nicidx += 1
yield aname, yieldinf
def _get_adp_urls(self):
adpurls = self.sysinfo.get('PCIeDevices', [])
if adpurls:
urls = [x['@odata.id'] for x in adpurls]
else:
urls = []
return urls
def _get_cpu_inventory(self, onlynames=False, withids=False, urls=None):
if not urls:
cpurl = self.sysinfo.get('Processors', {}).get('@odata.id', None)
if cpurl is None:
return
cpurl = self._do_web_request(cpurl)
urls = [x['@odata.id'] for x in cpurl.get('Members', [])]
urls = self._get_cpu_urls()
if not urls:
return
for res in self._do_bulk_requests(urls):
currcpuinfo, url = res
name = currcpuinfo.get('Name', 'CPU')
@ -624,13 +634,20 @@ class Command(object):
cpuinfo = {'Model': currcpuinfo.get('Model', None)}
yield (name, cpuinfo)
def _get_cpu_urls(self):
cpurl = self.sysinfo.get('Processors', {}).get('@odata.id', None)
if cpurl is None:
urls = []
else:
cpurl = self._do_web_request(cpurl)
urls = [x['@odata.id'] for x in cpurl.get('Members', [])]
return urls
def _get_mem_inventory(self, onlyname=False, withids=False, urls=None):
if not urls:
memurl = self.sysinfo.get('Memory', {}).get('@odata.id', None)
if not memurl:
return
memurl = self._do_web_request(memurl)
urls = [x['@odata.id'] for x in memurl.get('Members', [])]
urls = self._get_mem_urls()
if not urls:
return
for mem in self._do_bulk_requests(urls):
currmeminfo, url = mem
name = currmeminfo.get('Name', 'Memory')
@ -660,6 +677,15 @@ class Command(object):
}
yield (name, meminfo)
def _get_mem_urls(self):
memurl = self.sysinfo.get('Memory', {}).get('@odata.id', None)
if not memurl:
urls = []
else:
memurl = self._do_web_request(memurl)
urls = [x['@odata.id'] for x in memurl.get('Members', [])]
return urls
if __name__ == '__main__':
import os