| Bug #46490 | Full table scan hangs with more than 21 fragments | ||
|---|---|---|---|
| Submitted: | 31 Jul 2009 12:12 | Modified: | 6 Aug 2009 8:13 |
| Reporter: | Oli Sennhauser | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S2 (Serious) |
| Version: | mysql-5.1-telco-6.2 | OS: | Any |
| Assigned to: | Frazer Clement | CPU Architecture: | Any |
| Tags: | cluster, fragment, full, hang, scan, table | ||
[4 Aug 2009 10:38]
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/79988 2960 Frazer Clement 2009-08-04 Bug#46490 : Full table scan hangs with more than 21 fragments modified: storage/ndb/include/kernel/signaldata/ScanTab.hpp storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp
[4 Aug 2009 10:53]
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/79991 2960 Frazer Clement 2009-08-04 Bug#46490 : Full table scan hangs with more than 21 fragments modified: storage/ndb/include/kernel/signaldata/ScanTab.hpp storage/ndb/src/kernel/blocks/dbtc/DbtcMain.cpp
[5 Aug 2009 12:59]
Jonas Oreland
pushed to 6.2.19, 6.3.26 and 7.0.7
[6 Aug 2009 8:13]
Jon Stephens
Documented bugfix in the NDB-6.2.19, 6.3.26, and 7.0.7 changelogs as follows:
Full table scans failed with more than 21 fragments.
The number of fragments is the number of data nodes, times 8
(that is, MAX_FRAG_PER_NODE), divided by the number of replicas.
Thus, when NoOfReplicas = 1 at least 3 data nodes were required
to trigger this issue, and when NoOfReplicas = 2 at least 4 data
nodes were required to do so.
Description: MySQL cluster hangs on full table scans and gets error at the end: ndb_select_all cache_update -d vm_vs > cache_update.select ERROR: 4008 Receive from NDB failed Status: Unknown result, Classification: Unknown result error File: select_all.cpp (Line: 442) How to repeat: There seems to be a rule not to create > 4 fragments/node so you need > 4 nodes to reproduce Then put data in and try some scans Sooner or later should get a hang In debug, the 'unhandled signals after execute()' results in a node crash Suggested fix: Frazer is looking...