Bug #39549 Incorrect handled resource shortage of TC::scanfragreq causing node failure
Submitted: 19 Sep 2008 23:34 Modified: 20 Sep 2008 14:31
Reporter: Jonas Oreland
Status: Closed
Category:Server: Cluster Severity:S3 (Non-critical)
Version:* OS:Any
Assigned to: Jonas Oreland Target Version:

[19 Sep 2008 23:34] Jonas Oreland 
Description:
If setting a low value of MaxNoOfLocalScans and performing lots of (certain) scans. TC
could run out of scanfragrecords, and incorrectly crash instead of
returning an appropriate error code

How to repeat:
MaxNoOfLocalScan=32
2-node cluster
table with 10000 rows,
perform 100 parallel "select * from t1 where a > 9997 order by a for update"
in endless loop

Suggested fix:
I introduced a new error code 291,
with the error message
"Out of scanfrag records in TC (increase MaxNoOfLocalScans)
[19 Sep 2008 23:46] Bugs System 
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/54376

2669 Jonas Oreland	2008-09-19
      ndb - bug#39549 - incorrectly handled resource shortage in TC::scanfragrec
[20 Sep 2008 0:08] Bugs System 
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/54381

2669 Jonas Oreland	2008-09-19
      ndb - bug#39549 - incorrectly handled resource shortage in TC::scanfragrec
[20 Sep 2008 0:16] Jonas Oreland 
pushed to 62 63 and 64
[20 Sep 2008 14:31] Jon Stephens 
Documented in the NDB 6.2.16 and 6.3.18 changelogs as follows:

        Setting a low value of MaxNoOfLocalScans (< 100) and performing a
        large number of (certain) scans could cause the Transaction Coordinator
        to run out of scan fragment records, and then crash. Now when this
        resource is exhausted, the cluster returns Error 291 (Out of scanfrag
        records in TC (increase MaxNoOfLocalScans)) instead.
[25 Sep 2008 12:57] Bugs System 
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/54511

2671 Jonas Oreland	2008-09-25
      ndb - atrt
        fix replication in 6.0
        allow configurations wo/ cluster
[13 Dec 2008 0:28] Bugs System 
Pushed into 6.0.7-alpha  (revid:jonas@mysql.com-20080919214900-cjrhehof2t2pjsnf) (version
source revid:jonas@mysql.com-20080925105539-wd6gbofp5alv9j93) (pib:5)