Bug #62650 | Data Node crashes in DBDIH during node restart (followed by restore of mysqldum) | ||
---|---|---|---|
Submitted: | 7 Oct 2011 8:12 | Modified: | 7 Nov 2011 11:02 |
Reporter: | Johan Andersson | Email Updates: | |
Status: | No Feedback | Impact on me: | |
Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S3 (Non-critical) |
Version: | 7.1.15a | OS: | Any |
Assigned to: | CPU Architecture: | Any | |
Tags: | dbdih, Failure, ndbmtd, node restart |
[7 Oct 2011 8:12]
Johan Andersson
[7 Oct 2011 8:14]
Johan Andersson
I would love to ftp tracefiles, but ftp.mysql.com is not working.
[7 Oct 2011 8:25]
Johan Andersson
file is split using 'split' - 4 parts in total
Attachment: xaa (application/octet-stream, text), 500.00 KiB.
[7 Oct 2011 8:26]
Johan Andersson
file is split using 'split' - 4 parts in total
Attachment: xac (application/octet-stream, text), 500.00 KiB.
[7 Oct 2011 8:26]
Johan Andersson
file is split using 'split' - 4 parts in total
Attachment: xab (application/octet-stream, text), 500.00 KiB.
[7 Oct 2011 8:27]
Johan Andersson
file is split using 'split' - 4 parts in total - original filename: ndb_error_report_20111005013312.tar.bz2
Attachment: xad (application/octet-stream, text), 216.64 KiB.
[7 Oct 2011 11:00]
Jonas Oreland
Hi Johan, It's LCP related. And (current) LCP is as everyone knows fubar. But your case looks like a corner case. Any attaching a quick and dirty maybe fix (i didn't manage to reproduce your problem) /Jonas
[7 Oct 2011 11:01]
Jonas Oreland
maybe fix
Attachment: johan.patch (text/x-patch), 430 bytes.
[7 Oct 2011 11:01]
Jonas Oreland
btw, was the crash reproducible ?
[7 Oct 2011 12:07]
Johan Andersson
Thanks for the incredibly fast turnaround. Yes, reproducable every time. Does the fix cover both the DBDIH and the LCP problem?
[7 Oct 2011 12:14]
Jonas Oreland
1) hold on with the blessing until we see if it fixes problem 2) pretty sure the LCP is caused by first problem (although not 100%) /Jonas
[9 Nov 2011 11:08]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".