09ead69096
Currently there is an 'all(fragments)' check in the ECAppIter to know when the iter is done. It assumes that all the fragments will be emtpy when they are done. But if this is not the case we should log. We have no idea if how often this can happen but currently we don't log it so have no idea. But it could be a cause of some EC short read 500 that clusters get from time to time. Here is an example of the message: May 26 23:42:57 saio proxy-server[7942]: Un-recoverable fragment rebuild. Only received 3/4 fragments for '/a/c/o' (txn: txb43f239d78834bdabcaa6-0060aedd01) (client_ip: 127.0.0.1) Change-Id: If68e72d613685081a3149aedbfd07a69b8669d7f |
||
|---|---|---|
| .. | ||
| account | ||
| cli | ||
| common | ||
| container | ||
| obj | ||
| proxy | ||
| test_locale | ||
| __init__.py | ||
| helpers.py | ||