[jira] [Commented] (TS-2259) Introduce failover for logging system

["Yunkai Zhang (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TS-2259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13781335#comment-13781335 ] 

Yunkai Zhang commented on TS-2259:
----------------------------------

We had discussed this feature in email:

==James Peach==
Hi Yunkai,

We have been asked to look at ATS log host failover, so I started reading the logging source code. It looks to me that I could easily modify LogHostList::preproc_and_try_delete() to just write to the first available log host. After that, I could add more advanced logic to keep the current host if a failed log host came back to life ....

Do you think that this is a reasonable approach?

==Yunkai Zhang==
"just write to the first available log host" seems will break something -- user may want to write the log to multiple log hosts at the same time.

Maybe we can add *m_next_failover_host* member into LogHost:

class LogHost {
   ...
   LogHost *m_next_failover_host;
   ...
}; 

And then, in LogHostList::preproc_and_try_delete(), make the for-loop looks like:

for (LogHost * host = first(); host && nr; host = next(host)) {
  int ret = host->preproc_and_try_delete(lb);
  while (ret == FAILOVER) {
    LogHost *failover_host = host->m_next_failover_host;

    if (failover_host == NULL) //Fail to failover
      break;

    ret = failover_host->preproc_and_try_delete(lb);
  }
  nr--;
} 

FYI:)

==Yongming Zhao==
I'd like the idea of adding a failover log host option, that will be a good enhancement.

+1

> Introduce failover for logging system
> -------------------------------------
>
>                 Key: TS-2259
>                 URL: https://issues.apache.org/jira/browse/TS-2259
>             Project: Traffic Server
>          Issue Type: New Feature
>          Components: Logging
>            Reporter: Yunkai Zhang
>




--
This message was sent by Atlassian JIRA
(v6.1#6144)