[SERVER-46024] Collect /proc/vmstat swap statistics in FTDC Created: 06/Feb/20 Updated: 29/Oct/23 Resolved: 18/Mar/20 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Diagnostics |
| Affects Version/s: | None |
| Fix Version/s: | 4.2.6, 3.6.18, 4.4.0-rc0, 4.0.18, 4.7.0 |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Kevin Arhelger | Assignee: | Mark Benvenuto |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Backwards Compatibility: | Fully Compatible | ||||||||
| Backport Requested: |
v4.4, v4.2, v4.0, v3.6
|
||||||||
| Sprint: | Security 2020-03-23 | ||||||||
| Participants: | |||||||||
| Description |
|
FTDC currently collects page faults and disk activity, however determining the status of actual swap activity is difficult from just these metrics. The addition of pswpin and pswpout should cover this activity. |
| Comments |
| Comment by Githook User [ 25/Mar/20 ] |
|
Author: {'email': 'mark.benvenuto@mongodb.com', 'name': 'Mark Benvenuto', 'username': 'markbenvenuto'}Message: (cherry picked from commit 97e0731610f18576229c322c3a4186202c77c521) |
| Comment by Githook User [ 20/Mar/20 ] |
|
Author: {'name': 'Mark Benvenuto', 'username': 'markbenvenuto', 'email': 'mark.benvenuto@mongodb.com'}Message: (cherry picked from commit 97e0731610f18576229c322c3a4186202c77c521) |
| Comment by Githook User [ 19/Mar/20 ] |
|
Author: {'email': 'mark.benvenuto@mongodb.com', 'name': 'Mark Benvenuto', 'username': 'markbenvenuto'}Message: (cherry picked from commit 97e0731610f18576229c322c3a4186202c77c521) |
| Comment by Githook User [ 18/Mar/20 ] |
|
Author: {'name': 'Mark Benvenuto', 'username': 'markbenvenuto', 'email': 'mark.benvenuto@mongodb.com'}Message: (cherry picked from commit 97e0731610f18576229c322c3a4186202c77c521) |
| Comment by Githook User [ 17/Mar/20 ] |
|
Author: {'name': 'Mark Benvenuto', 'username': 'markbenvenuto', 'email': 'mark.benvenuto@mongodb.com'}Message: |
| Comment by Mark Benvenuto [ 11/Mar/20 ] |
|
Yes, you are correct. The pgmajfault counter faults that require reading from disk and pgfault is all faults (aka "soft" or "minor"). I will add the pgfault counter. |
| Comment by Bruce Lucas (Inactive) [ 11/Mar/20 ] |
|
Is the difference between pgmajfault and pgfault that the latter includes "soft" or "minor" page faults used to implement on-demand memory allocation? If so this might be interesting because I've at times suspected that a high rate of minor page faults due to allocator commit/decommit churn was responsible for some performance issues. |
| Comment by Mark Benvenuto [ 11/Mar/20 ] |
|
I would also include nr_mlock. bruce.lucas, what are your thoughts? In summary,
Are the counters I am planning on including.
References https://elixir.bootlin.com/linux/latest/source/include/linux/vm_event_item.h#L25 |
| Comment by Mark Benvenuto [ 10/Mar/20 ] |
|
Should we also include pgmajfault for major page faults? I could not find any counters that really stand out: pgfault is all page faults which is not interesting. |