"Fossies" - the Fresh Open Source Software Archive  

Source code changes of the file "docs/reference/elastic-search.md" between
wire-server-2020-06-10.tar.gz and wire-server-2020-06-19.tar.gz

About: Wire (server) offers end-to-end encrypted messaging, file-sharing, video and voice calls, and guest rooms for external communication (back-end server).

elastic-search.md  (wire-server-2020-06-10):elastic-search.md  (wire-server-2020-06-19)
skipping to change at line 115 skipping to change at line 115
docker run "quay.io/wire/brig-index:$WIRE_VERSION" reindex-from-another-index \ docker run "quay.io/wire/brig-index:$WIRE_VERSION" reindex-from-another-index \
--elasticsearch-server "http://$ES_HOST:$ES_PORT" \ --elasticsearch-server "http://$ES_HOST:$ES_PORT" \
--source-index "$ES_SRC_INDEX" \ --source-index "$ES_SRC_INDEX" \
--destination-index "$ES_DEST_INDEX" --destination-index "$ES_DEST_INDEX"
``` ```
Optionally, `--timeout <NUMBER_OF_SECONDS>` can be added to increase/decrease from the default timeout of 10 minutes. Optionally, `--timeout <NUMBER_OF_SECONDS>` can be added to increase/decrease from the default timeout of 10 minutes.
1. Redeploy brig without `elasticsearch.additionalWriteIndex` and with `elastics earch.index` set to the name of new index 1. Redeploy brig without `elasticsearch.additionalWriteIndex` and with `elastics earch.index` set to the name of new index
Now you can delete the old index. Now you can delete the old index.
**NOTE**: There is a bug hidden when using this way. Sometimes a user won't get
deleted from the index. Attempts at reproducing this issue in a simpler
environment have failed. As a workaround, there is a tool in
[tools/db/find-undead](../../tools/db/find-undead) which can be used to find the
undead users right after the migration. If they exist, please run refill the ES
documents from cassandra as described [above](#refill-es-documents-from-cassandr
a)
## Recreate an index (Requires downtime) ## Recreate an index (Requires downtime)
When analysis settings of an index need to be changed, e.g. for changes When analysis settings of an index need to be changed, e.g. for changes
introduced in [#1052](https://github.com/wireapp/wire-server/pull/1052), introduced in [#1052](https://github.com/wireapp/wire-server/pull/1052),
it is not possible to keep the index running while the changes are applied. it is not possible to keep the index running while the changes are applied.
To tackle this, a wire-server operator must either migrate to a new index as To tackle this, a wire-server operator must either migrate to a new index as
documented [above](#migrate-to-a-new-index) or allow for some downtime. One documented [above](#migrate-to-a-new-index) or allow for some downtime. One
might want to choose downtime for simplicity. These steps are especially simple might want to choose downtime for simplicity. These steps are especially simple
to do when using [wire-server-deploy](https://github.com/wireapp/wire-server-dep loy/). to do when using [wire-server-deploy](https://github.com/wireapp/wire-server-dep loy/).
 End of changes. 1 change blocks. 
0 lines changed or deleted 8 lines changed or added

Home  |  About  |  Features  |  All  |  Newest  |  Dox  |  Diffs  |  RSS Feeds  |  Screenshots  |  Comments  |  Imprint  |  Privacy  |  HTTP(S)