Bug #55622 | ndbmtd node start causes all cluster to fail | ||
---|---|---|---|
Submitted: | 29 Jul 2010 9:48 | Modified: | 8 Sep 2016 6:45 |
Reporter: | Alexey Asemov | Email Updates: | |
Status: | Unsupported | Impact on me: | |
Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S1 (Critical) |
Version: | mysql-5.1-telco-7.1 | OS: | Linux (CentOS 5.5) |
Assigned to: | CPU Architecture: | Any | |
Tags: | mysql-5.1.44 ndb-7.1.4b |
[29 Jul 2010 9:48]
Alexey Asemov
[29 Jul 2010 10:08]
Alexey Asemov
bug-data-55622.zip uploaded to the bug data upload FTP, it contains all the tracebacks.
[10 Aug 2010 13:17]
Hartmut Holzgraefe
Running with NoOfReplicas=3 or 4 is not well tested and is known to have bugs, we currently only actually support NoOfReplicas=1 (for pure speed, no HA) and =2 (for HA) ...
[16 Aug 2010 10:59]
Hartmut Holzgraefe
Can you reproduce the same with NoOfReplicas=2 ?
[26 Aug 2010 7:01]
Alexey Asemov
Transitioning to two-replica ndbmtd cluster got me with plenty of LCP failures with NDBFS error code 1217. Transitioning to ndbd solved LCP failures, and is now running smooth. I will try a bit of node restarts in the meanwhile and report accordingly.
[24 Sep 2010 7:29]
Alexey Asemov
May be related to #54945? Can I get 7.1.8 sources for test?
[8 Sep 2016 6:45]
MySQL Verification Team
> 2 replica's is still unsupported. It's experimental still with latest versions of the ndbcluster. What is reported here is still more/less expected behavior. kind regards Bogdan Kecman