[SERVER-20459] pingMs and lastHearbeatMessage fields are sporadic Created: 17/Sep/15 Updated: 07/Oct/15 Resolved: 21/Sep/15 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Diagnostics |
| Affects Version/s: | None |
| Fix Version/s: | 3.1.9 |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Bruce Lucas (Inactive) | Assignee: | Eric Milkie |
| Resolution: | Done | Votes: | 0 |
| Labels: | FTDC | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Backwards Compatibility: | Fully Compatible |
| Operating System: | ALL |
| Sprint: | RPL A (10/09/15) |
| Participants: |
| Description |
|
The pingMs and lastHearbeatMessage fields in the members array of the replSetGetStatus command occur sporadically. The fields appear to be alternately present for a couple of seconds and missing for a couple of seconds. This creates frequent schema changes that reduce the compression efficiency of FTDC storage by a factor of 5-10x. Observed behavior in 3.0 is that pingMs is always present (and lastHearbeatMessage seems to be new for 3.1). |
| Comments |
| Comment by Githook User [ 21/Sep/15 ] |
|
Author: {u'username': u'milkie', u'name': u'Eric Milkie', u'email': u'milkie@10gen.com'}Message: |
| Comment by Bruce Lucas (Inactive) [ 17/Sep/15 ] |
|
It would have been significant effort to set up a cluster with a ping time consistently >1ms, so instead I tried the proposed patch and it (or equivalent) would fix the problem, thanks. |
| Comment by Eric Milkie [ 17/Sep/15 ] |
|
I believe the new behavior is that if the recent ping time happens to be less than 1ms, it rounds to 0 and then doesn't display the pingMs field. The old behavior in 3.0 was to display the field even if the ping time was 0. |