jackrabbit-oak-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrei Dulceanu (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (OAK-6673) Improve cold standby logging
Date Fri, 22 Sep 2017 14:45:00 GMT

     [ https://issues.apache.org/jira/browse/OAK-6673?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Andrei Dulceanu updated OAK-6673:
    Attachment: OAK-6673.patch

I implemented all the changes in the patch attached. I'm not really happy with the solution
found for tracking chunks/overall chunk size used on the server in {{ResponseDecoder}}, but
couldn't find a better one. At least the log output is now clearer:
17:26:22.113 DEBUG [primary-1] ChunkedBlobStream.java:128   Sending chunk 122/128 of size
8388608 from blob 8098b6ac1491be80b7e58a85767ede178c432866d90caf6726f556406ecc84a4#1073741824
to client c11a6d85-16b9-4ba1-b4c5-8ec9129dac33
17:26:22.179 DEBUG [standby-1] ResponseDecoder.java:90      Decoding 'get blob' response
17:26:22.179 DEBUG [standby-1] ResponseDecoder.java:150     Received chunk 122/128 of size
8388608 from blob 8098b6ac1491be80b7e58a85767ede178c432866d90caf6726f556406ecc84a4#1073741824
17:26:22.190 DEBUG [standby-1] ResponseDecoder.java:159     All checks OK. Appending chunk
to disk to /var/folders/1s/pdyvv3kd33v2x1_fckz4v2pr0000gn/T/8098b6ac1491be80b7e58a85767ede178c432866d90caf6726f556406ecc84a4#1073741824.tmp


[~frm], could you take a look at the patch, please?

> Improve cold standby logging
> ----------------------------
>                 Key: OAK-6673
>                 URL: https://issues.apache.org/jira/browse/OAK-6673
>             Project: Jackrabbit Oak
>          Issue Type: Task
>          Components: segment-tar, tarmk-standby
>            Reporter: Andrei Dulceanu
>            Assignee: Andrei Dulceanu
>            Priority: Minor
>              Labels: cold-standby
>             Fix For: 1.7.8
>         Attachments: OAK-6673.patch
> With the new feature which allows chunking in blob transfer between server and client,
there are various places in which more meaningful log messages could be used. For example,
on the server, there's a tally of the no. of chunks sent/total no. of chunks, but this part
is missing on the client, i.e. no. of chunks received/total no. of chunks. 
> Another case which would benefit from improved logging is when a big blob can't be sent
fully from the server to the client in {{readTimeoutMs}}. The current exception message is
a bit scarce in details (e.g. {{"Unable to load remote blob " + blobId + " at " + path + "#"
+ pName}}). This could also mean that the remote blob doesn't exist on the server in the first
place. A better option would be to advise about increasing {{readTimeoutMs}}.
> Finally, the same log level should be used everywhere, since currently {{DEBUG}} and
{{INFO}} are  interchangeably mixed.
> /cc [~frm]

This message was sent by Atlassian JIRA

View raw message