You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-dev@lucene.apache.org by "Bryan Talbot (JIRA)" <ji...@apache.org> on 2009/05/15 18:19:45 UTC
[jira] Created: (SOLR-1170) Java replication replicates lucene lock
file
Java replication replicates lucene lock file
--------------------------------------------
Key: SOLR-1170
URL: https://issues.apache.org/jira/browse/SOLR-1170
Project: Solr
Issue Type: Bug
Components: replication (java)
Affects Versions: 1.4
Environment: linux
Reporter: Bryan Talbot
Priority: Minor
When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
Why should the lock file be replicated to slaves?
The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710716#action_12710716 ]
Noble Paul commented on SOLR-1170:
----------------------------------
are you sure that the lucene-write.lock is actually replicated . It may be created in the slave itself. to confirm it can your check the master with the command=filelist&indexversion=<the version>
the version can be obtained from the command=indexversion
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (SOLR-1170) Java replication replicates lucene lock
file
Posted by "Hoss Man (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Hoss Man updated SOLR-1170:
---------------------------
Fix Version/s: 1.4
issue is marked "affects 1.4" and has a patch which comments indicate solves the problem ... marking "fix for 1.4" so it's not overlooked prior to code freeze
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Fix For: 1.4
>
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Noble Paul resolved SOLR-1170.
------------------------------
Resolution: Fixed
committed r815597
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Fix For: 1.4
>
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Bryan Talbot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12711015#action_12711015 ]
Bryan Talbot commented on SOLR-1170:
------------------------------------
When the slave is cleanly shutdown, the lock file remains. Doesn't that indicate an index writer got left open and not properly shutdown?
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Mark Miller (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710775#action_12710775 ]
Mark Miller commented on SOLR-1170:
-----------------------------------
The replication handler calls forceOpenWriter to ensure old index files are deleted after a pull.
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12754498#action_12754498 ]
Noble Paul commented on SOLR-1170:
----------------------------------
I guess it safe enough to commit this patch. The .lock file does not need to be replicated in any case
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Fix For: 1.4
>
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12711034#action_12711034 ]
Noble Paul commented on SOLR-1170:
----------------------------------
even if the lock is present it can be removed if the attribute mainIndex/unlockOnStartup is present in the solrconfig.xml.
But if it does not get removed automatically there could be a problem
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Noble Paul reassigned SOLR-1170:
--------------------------------
Assignee: Noble Paul
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Bryan Talbot (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710767#action_12710767 ]
Bryan Talbot commented on SOLR-1170:
------------------------------------
The lock file doesn't show up in the filelist. It always has the same timestamp as other files which were replicated and sticks around until the next replication begins so I had assumed it was replicated.
If it is created on the slave instead of being replicated, isn't that still an issue? I'm certain that there are no writes being performed on the slave.
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (SOLR-1170) Java replication replicates lucene lock
file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Noble Paul updated SOLR-1170:
-----------------------------
Attachment: SOLR-1170.patch
if it is indeed a problem. Then this patch can solve it
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (SOLR-1170) Java replication replicates lucene
lock file
Posted by "Noble Paul (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/SOLR-1170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12711012#action_12711012 ]
Noble Paul commented on SOLR-1170:
----------------------------------
in that case, if you promote your slave to become a master it should have no problem. the lock is created by the same server ,so it will be removed when you close the writer or restart the server
> Java replication replicates lucene lock file
> --------------------------------------------
>
> Key: SOLR-1170
> URL: https://issues.apache.org/jira/browse/SOLR-1170
> Project: Solr
> Issue Type: Bug
> Components: replication (java)
> Affects Versions: 1.4
> Environment: linux
> Reporter: Bryan Talbot
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-1170.patch
>
>
> When using solr 1.4 replication, I see that the lucene-write.lock file is being replicated to slaves. I'm importing data from a db every 5 minutes using cron to trigger a DIH delta-import. Replication polls every 60 seconds and the master is configured to take a snapshot (replicateAfter) commit.
> Why should the lock file be replicated to slaves?
> The lock file isn't stale on the master and is absent unless the delta-import is in process. I've not tried it yet, but with the lock file replicated, it seems like promotion of a slave to a master in a failure recovery scenario requires the manual removal of the lock file.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.