[GitHub] nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads.

nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads. URL: https://github.com/apache/bookkeeper/pull/1804#issuecomment-438063537 Sounds good to me. The most common case is probably addressed by the earlier change.

[GitHub] nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads.

nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads. URL: https://github.com/apache/bookkeeper/pull/1804#issuecomment-438056753 I *think* what we need is something like this: if (isHandleWritable() || metadata.isInRe

[GitHub] nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads.

nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads. URL: https://github.com/apache/bookkeeper/pull/1804#issuecomment-438045377 During recovery I suppose they should... In the absence of a recovery on a read-only LH, they don't need

[GitHub] nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads.

nicmichael commented on issue #1804: [client] read submission for readonly handle should bypass OSE threads. URL: https://github.com/apache/bookkeeper/pull/1804#issuecomment-437985501 @sijie Yes that was the intention. I'm looking at the test failures now. Adding "|| isRecoveryRead" to