MySQL Bugs: #92284: MySQL NDB Cluster: 7.5.10: Restart hangs at index rebuild for some LDM

Bug #92284	MySQL NDB Cluster: 7.5.10: Restart hangs at index rebuild for some LDM
Submitted:	4 Sep 2018 12:36	Modified:	9 Sep 2018 19:16
Reporter:	Thomas Waibel-BGo	Email Updates:
Status:	Not a Bug	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S2 (Serious)
Version:	7.5.10	OS:	Debian (9)
Assigned to:	MySQL Verification Team	CPU Architecture:	x86 (Intel(R) Xeon(R) CPU E5-2643 v2 @ 3.50GHz)

Description:
Hi,

after one of the 4 data nodes stopped the restart seems to hang at the index rebuild for some of the 12 LDMs.

5 of the LDMs get to the status of "We have completed restoring our fragments and executed REDO log and rebuilt ordered indexes" where as the others do not even start rebuilding indices.

The current state is that all CPUs are idle on that system an there is no further log output. This is the second restart after the initial stop.

Please find the ndb error report here
https://data.boerse-go.de/it-operations/ndb_error_report_20180904141958.tar.bz2

Best regards

How to repeat:
There was a node stop because of 
Node 3 declared dead due to missed heartbeat

The Restart of the node is just stopping with last status at
NR Status: node=3,OLD=Undo Disk data ongoing,NEW=Execute REDO logs ongoing

Suggested fix:
There is no proper log output suggesting if or where a problem is occuring.

Hi,

This does not look like a bug but miss-configuration / improper sizing of the cluster. You might want to contact tech support.

kind regards