Changeset a0a770a in trunk
- Timestamp:
- 2010-05-24T04:28:36Z (15 years ago)
- Branches:
- master
- Children:
- 7cadb49b
- Parents:
- 4141d95
- Location:
- docs
- Files:
-
- 2 edited
Legend:
- Unmodified
- Added
- Removed
-
TabularUnified docs/architecture.txt ¶
r4141d95 ra0a770a 196 196 with as many shares as we desire. 197 197 198 If we are unable to place every share that we want, but we still managed to 199 place enough shares on enough servers to achieve a condition called "servers of 200 happiness" then we'll do the upload anyways. If we cannot achieve "servers of 201 happiness", the upload is declared a failure. 198 Before a file upload is called successful, it has to pass an upload health 199 check. For immutable files, we check to see that a condition called 200 'servers-of-happiness' is satisfied. When satisfied, 'servers-of-happiness' 201 assures us that enough pieces of the file are distributed across enough 202 servers on the grid to ensure that the availability of the file will not be 203 affected if a few of those servers later fail. For mutable files and 204 directories, we check to see that all of the encoded shares generated during 205 the upload process were successfully placed on the grid. This is a weaker 206 check than 'servers-of-happiness'; it does not consider any information about 207 how the encoded shares are placed on the grid, and cannot detect situations in 208 which all or a majority of the encoded shares generated during the upload 209 process reside on only one storage server. We hope to extend 210 'servers-of-happiness' to mutable files in a future release of Tahoe-LAFS. If, 211 at the end of the upload process, the appropriate upload health check fails, 212 the upload is considered a failure. 202 213 203 214 The current defaults use k=3, servers_of_happiness=7, and N=10. N=10 means that 204 215 we'll try to place 10 shares. k=3 means that we need any three shares to 205 recover the file. servers_of_happiness=7 means that we'll consider the upload206 to be successful if we can place shares on enough servers that there are 7 207 different servers, the correct functioning of any k of which guarantee the 208 availability of the file.216 recover the file. servers_of_happiness=7 means that we'll consider an immutable 217 file upload to be successful if we can place shares on enough servers that 218 there are 7 different servers, the correct functioning of any k of which 219 guarantee the availability of the immutable file. 209 220 210 221 N=10 and k=3 means there is a 3.3x expansion factor. On a small grid, you -
TabularUnified docs/configuration.txt ¶
r4141d95 ra0a770a 273 273 uses. 274 274 275 shares.happy allows you control over the distribution of your file. An upload 276 is only considered successful if shares are placed on at least 'shares.happy' 277 distinct servers, the correct functioning of at least k of which is sufficient 278 to guarantee the availability of the uploaded file. This value should not be 279 larger than the number of servers on your grid. 275 shares.happy allows you control over the distribution of your immutable file. 276 An upload is only considered successful if shares are placed on at least 277 'shares.happy' distinct servers, the correct functioning of at least k of 278 which is sufficient to guarantee the availability of the uploaded file. This 279 value should not be larger than the number of servers on your grid. 280 281 (Mutable files use a different share placement algorithm that does not 282 consider this parameter.) 280 283 281 284
Note: See TracChangeset
for help on using the changeset viewer.