| Bug #20823 | testBackup FailMaster failing | ||
|---|---|---|---|
| Submitted: | 3 Jul 2006 14:15 | Modified: | 1 Sep 2006 8:02 |
| Reporter: | Stewart Smith | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S2 (Serious) |
| Version: | 5.1 | OS: | |
| Assigned to: | Stewart Smith | CPU Architecture: | Any |
[3 Jul 2006 14:15]
Stewart Smith
[6 Jul 2006 13:23]
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/8825
[7 Jul 2006 7:25]
Jonas Oreland
Question: wouldnt it be "smaller" an more obvious to instead add an if-statement
in Backup::sendDropTrig
And would it work...i.e is the tables array always cleared
after all files are closed...
(my guess would be yes...but you have to check)
Note: When this patch is ready it has to be backported to 5.0 aswell...
[2 Aug 2006 8:41]
Stewart Smith
the addition also includes:
if(tabPtr.p->fragments.getSize())
{
// do foo
as it's possible to have a valid tabPtr but not have any information on the fragments (hence 0 length tabPtr.p->fragments)
which is possible if crash between GSN_DI_FCOUNTREQ and DI_FCOUNTCONF
[8 Aug 2006 6:56]
Stewart Smith
<jonas> stewart: does you comment mean: no jonas suggestion does not work ? <stewart> jonas: it would work, but would require dereferencing the Ptr twice in the common (non failure) case. I think this way is more optimal in the common case. <jonas> stewart: note: I dont really remember details but...I think making code more obvious and dereferencing Ptr twice in dropdrig is a fair trade <jonas> stewart: note2: if this was in "transid_ai" i would reason differently... <stewart> jonas: will fix (and add this transcript to the bug report so i remember).
[10 Aug 2006 6:27]
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/10240 ChangeSet@1.2228, 2006-08-10 14:27:00+08:00, stewart@willster.(none) +1 -0 BUG#20823 testBackup failing in FailMaster fix after jonas review
[15 Aug 2006 3:10]
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/10397 ChangeSet@1.2246, 2006-08-15 11:09:38+08:00, stewart@willster.(none) +1 -0 BUG#20823 testBackup FailMaster failing few cases not handled properly (NF occurs).
[15 Aug 2006 5:07]
Stewart Smith
pushed to 5.0-ndb team tree.
[1 Sep 2006 8:02]
Jonas Oreland
pushed to 5.1.12, nothing to document
