Skip to content

Revconn refresh fixes - #6209

Open
markhannum wants to merge 2 commits into
bloomberg:mainfrom
markhannum:revconn-refresh-fixes
Open

markhannum wants to merge 2 commits into
bloomberg:mainfrom
markhannum:revconn-refresh-fixes

Conversation

@markhannum

@markhannum markhannum commented Sep 11, 2026

Copy link
Copy Markdown
Contributor

A metadb that failed to answer a reverse-connection refresh was indistinguishable from one that answered with nothing, so a single unreachable metadb marked live workers EXITING and tore down healthy reverse connections. Each host is now attributed to the metadbs that vouched for it and is retired only once all of them have answered and dropped it. Traces name the metadb, and stat dumprevsql reports the mask.

Also fixes replace_tier_by_hostname() -- bad unlink, unsafe iteration, leaked handle -- and two strtok_r() calls that truncated gbl_physrep_source_host/gbl_physrep_metadb_host in place at the first comma.

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
analyze_partial_index_off_generated [failed with core dumped] **quarantined**
analyze [failed with core dumped] **quarantined**
sc_resume_logicalsc_generated **quarantined**
tsa
consumer_non_atomic_default_consumer_generated **quarantined**
tunables
sc_downgrade [timeout] **quarantined**
sc_truncate_lockorder_generated [timeout] **quarantined**
skipscan [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 446517c to 6f09081 Compare September 11, 2026 19:39

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
analyze_partial_index_off_generated [failed with core dumped] **quarantined**
analyze [failed with core dumped] **quarantined**
tsa
ssl_san
consumer_non_atomic_default_consumer_generated **quarantined**
tunables
ssl_set_cmd
ssl_dbname
ssl_prefer
sc_downgrade [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 6f09081 to 0e7a1ad Compare September 12, 2026 09:07

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
sc_resume_logicalsc_generated **quarantined**
reco-ddlk-sql **quarantined**
tsa
ssl_san
consumer_non_atomic_default_consumer_generated **quarantined**
ssl_set_cmd
ssl_dbname
ssl_prefer
sc_downgrade [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 0e7a1ad to 165ec07 Compare September 12, 2026 12:19

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
sc_truncate_multiddl_generated [db unavailable at finish] **quarantined**
sc_resume_logicalsc_generated **quarantined**
comdb2sys_pagesize_generated
tsa
ssl_san
consumer_non_atomic_default_consumer_generated **quarantined**
ssl_set_cmd
ssl_prefer
ssl_dbname
sc_downgrade [timeout] **quarantined**

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
timepart_retro
tsa
consumer_non_atomic_default_consumer_generated **quarantined**
bind_query_plan
sc_downgrade [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 165ec07 to 249b9d6 Compare September 14, 2026 17:53

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
timepart_retro
tsa
ssl_san
sql_logfill_autodisable
consumer_non_atomic_default_consumer_generated **quarantined**
ssl_set_cmd
ssl_prefer
ssl_dbname
sc_downgrade [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 249b9d6 to 89acc1c Compare September 15, 2026 15:46

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**
tsa [timeout]

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 89acc1c to 258c30d Compare September 16, 2026 12:43

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
comdb2sys [db unavailable at finish] **quarantined**
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**

@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 258c30d to 20e7824 Compare September 16, 2026 15:48

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Error ⚠.
Regression testing: Success ✓.

The first 10 failing tests are:
timepart_retro
disttxn
sc_resume_logicalsc_generated **quarantined**
comdb2sys_queueodh_generated
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**

Attribute each revconn host to the metadbs that vouched for it and
retire it only when every voucher has answered and dropped it, so one
unreachable metadb cannot stop pruning.  Fixes replace_tier_by_hostname.

Signed-off-by: Mark Hannum <mhannum@bloomberg.net>
Adds physrep_revconn_altmetadb_conn_error and
physrep_revconn_metadb_provenance.

Signed-off-by: Mark Hannum <mhannum@bloomberg.net>
@markhannum
markhannum force-pushed the revconn-refresh-fixes branch from 20e7824 to 7158739 Compare September 17, 2026 11:03

@roborivers roborivers left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cbuild submission: Success ✓.
Regression testing: Success ✓.

The first 10 failing tests are:
consumer_non_atomic_default_consumer_generated **quarantined**
sc_downgrade [timeout] **quarantined**

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants