Red Hat Bugzilla – Bug 1298615
Host page provides misleading data for hosts listed there
Last modified: 2016-07-20 09:23:34 EDT
Description of problem ====================== Information provided for each host listed on Hosts page are misleading and not based on reality. Version-Release number of selected component ============================================ # rpm -qa | grep rhsc rhscon-ui-0.0.6-0.1.alpha1.el7.noarch rhscon-core-0.0.6-0.1.alpha1.el7.x86_64 rhscon-ceph-0.0.4-0.1.alpha1.el7.x86_64 How reproducible ================ 100% Steps to Reproduce ================== 1. Accept at least 2 hosts via web ui 2. Go to "Hosts" page 3. Make sure that all hosts are idle 4. Open terminal window for each host and tun top there 5. See data provided for each host on the "Hosts" page: - cpu load - memory usage - kernel version - version (this seems to be a ceph version) Actual results ============== Even though that all host machines were idle (cpu usage below few percent), Hosts page provided cpu usage values randomly reporting values such as 7%, 23% or even 70%. The same issue applies to memory usage. Reported kernel version doesn't match: I see 4.0.4-301.fc22.x86_64 while the actual kernel on the host machines is 3.10.0-327.el7.x86_64. Version reported is 0.94.3, while actual ceph package is ceph-0.94.3-3.el7cp.x86_64. Expected results ================ Host values are not random or hardcoded and reflects actual state of the machine.
Note: neither kernel nor ceph version is reported in the host list page as pointed out in verification comment in BZ 1296187
Checking with ============= rhscon-ui-0.0.48-1.el7scon.noarch rhscon-core-selinux-0.0.34-1.el7scon.noarch rhscon-ceph-0.0.33-1.el7scon.x86_64 rhscon-core-0.0.34-1.el7scon.x86_64 ceph-installer-1.0.14-1.el7scon.noarch ceph-ansible-1.0.5-28.el7scon.noarch Verification ============ On all machines of the cluster (all MON and OSD machines), I started stress process[1] to push cpu utilization to 100% and let it run for longer period of time so that RHSC 2.0 had enough time to notice this new state. After 15 minutes, I see that the cpu charts were updated: I see 100% value reported for all machines in the list. Similar check was done for memory[2], and after another 15 or 20 minutes, the charts were updated as well. >> VERIFIED [1] the actual command: `stress --cpu 2` (since my machines had 2 cpus) [2] command `stress --vm 1 --vm-bytes 1G ` (I didn't intended to use all memory) (That said, during testing of this BZ I noticed one peculiar issue, which has been reported as new BZ 1358270. I created new BZ instead of failing this one because my testing shows that the data are no longer random and actual values are reported - which is the main concern of this BZ.)