Migrate Replication Documentation to Manual from Wiki
(DOCS-260)
|
|
| Status: | Closed |
| Project: | Documentation |
| Component/s: | manual |
| Affects Version/s: | None |
| Fix Version/s: | v1.0, Server_Docs_20231030, Server_Docs_20231106, Server_Docs_20231105, Server_Docs_20231113 |
| Type: | Sub-task | Priority: | Major - P3 |
| Reporter: | Sam Kleinman (Inactive) | Assignee: | Andrew Leung (Inactive) |
| Resolution: | Done | Votes: | 0 |
| Labels: | redirect | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Participants: | |
| Days since reply: | 11 years, 32 weeks, 1 day ago |
| Description |
|
Old: http://www.mongodb.org/display/DOCS/Replica+Sets+slaveDelay |
| Comments |
| Comment by Sam Kleinman (Inactive) [ 10/Jul/12 ] |
|
These comments showed up here because I flipped the ticket number 276/267. In any case, I've updated the value to reflect the bitmask, but I'm leaving the usepowerof2sizes behavior to reflect what Kevin said. |
| Comment by Kevin Matulef [ 10/Jul/12 ] |
|
@Scott I think you're right about the system/user flags being bitmasks. It's confusing to the user, though. Wouldn't it make more sense for stats() to display the meaning of the set flags, rather than just the numeric value of the bitmask? If so, let's file a ticket for it. As for the second part, I'm not sure your comment makes sense. Why would it it use the 128 byte bucket rather than 64 (both powers of 2, greater than 48)? My understanding of this feature is that it only affects the way memory is allocated for new docs. Instead of an n-byte doc being given n*padding_factor bytes, it's given the nearest power of 2 greater than n. Then all the holes that get created by document deletion are all powers of 2, so easier to reuse. |
| Comment by Scott Hernandez (Inactive) [ 10/Jul/12 ] |
|
This comment if for the last commit. The power of 2 docs do not make sense to me at all, nor do the user/systemFlags descriptions. It is not the number of flags which are shows in the collections stats for system/userFlags, but the actual flag value, which is bitmask for each possible flag (bit): https://github.com/mongodb/mongo/blob/master/src/mongo/db/namespace_details.h#L188 The power of 2 stuff is related to which size hole in the free-list is used to use for the new doc allocation. freelist buckets = [32, 64, 128, 256 ...] Normally if you have a document which is 48 bytes then the algorithm will find a hole starting at 64 byte bucket, and move up the list of buckets till it finds a hole to put the doc in. But if the power of 2 is enabled then it will start looking at the 128 byte bucket and move up until it finds something. It should really be called nextPowerOf2Hole or something. Do these both make more sense? |
| Comment by auto [ 06/Jul/12 ] |
|
Author: {u'date': u'2012-07-06T14:14:42-07:00', u'email': u'samk@10gen.com', u'name': u'Sam Kleinman'}Message: |
| Comment by auto [ 05/Jul/12 ] |
|
Author: {u'date': u'2012-07-05T16:01:11-07:00', u'email': u'samk@10gen.com', u'name': u'Sam Kleinman'}Message: |
| Comment by auto [ 05/Jul/12 ] |
|
Author: {u'date': u'2012-07-05T15:49:44-07:00', u'email': u'samk@10gen.com', u'name': u'Sam Kleinman'}Message: |
| Comment by auto [ 05/Jul/12 ] |
|
Author: {u'date': u'2012-07-05T15:16:07-07:00', u'email': u'samk@10gen.com', u'name': u'Sam Kleinman'}Message: |
| Comment by auto [ 05/Jul/12 ] |
|
Author: {u'date': u'2012-07-05T14:45:58-07:00', u'email': u'samk@10gen.com', u'name': u'Sam Kleinman'}Message: |
| Comment by auto [ 05/Jul/12 ] |
|
Author: {u'date': u'2012-07-05T14:19:11-07:00', u'email': u'andrew.leung@10gen.com', u'name': u'Andrew Leung'}Message: updated Delayed and member priority entry. |
| Comment by Andrew Leung (Inactive) [ 05/Jul/12 ] |
|
updated command reference. ready for redirect. |