[
https://issues.apache.org/jira/browse/KUDU-1551?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Todd Lipcon updated KUDU-1551:
------------------------------
Target Version/s: 1.1.0
Labels: newbie (was: )
Priority: Critical (was: Major)
We managed to repro this on a test cluster as well. This happens if there is a
log segment which is preallocated but has no header written at all. eg:
{code}
[todd@vd0240 kudu]$ ls -lh
/data/15/kudu/wals/66f406b67630427ab0cf438b4a98eded.recovery/wal-000000001
-rw------- 1 kudu kudu 64M Oct 19 21:59
/data/15/kudu/wals/66f406b67630427ab0cf438b4a98eded.recovery/wal-000000001
[todd@vd0240 kudu]$ sudo -u kudu xxd
/data/15/kudu/wals/66f406b67630427ab0cf438b4a98eded.recovery/wal-000000001 |
grep -v '0000 0000 0000 0000 0000 0000 0000 0000'
[todd@vd0240 kudu]$
{code}
This might be a good second or third bug fix for a new contributor. Otherwise
hopefully someone can get to it for 1.1. As a workaround in the meantime, one
can probably just remove the empty log segment.
> kudu-tserver can not start up after experience node crash
> ---------------------------------------------------------
>
> Key: KUDU-1551
> URL: https://issues.apache.org/jira/browse/KUDU-1551
> Project: Kudu
> Issue Type: Bug
> Components: log
> Affects Versions: 1.0.1
> Reporter: zhangsong
> Priority: Critical
> Labels: newbie
>
> after experience node crash ,kudu-tserver can not restarted successfully,
> Fatal log :
> "I0804 18:36:04.109537 1909 log_util.cc:359] Successfully rebuilt footer for
> segment:
> /export/servers/kudu/tserver_wal_data_7052/wals/2c329cd1b4bf4850b9b9e2e0613f73d9.recovery/wal-000000001
> (valid entries through byte offset 1993)
> I0804 18:36:04.109566 1909 tablet.cc:854] T 2c329cd1b4bf4850b9b9e2e0613f73d9
> Rewinding schema during bootstrap to Schema [
>
> ]
> F0804 18:36:04.109634 1909 memrowset.cc:102] Check failed:
> schema.has_column_ids()"
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
