Back to bug 1031687
| Who | When | What | Removed | Added |
|---|---|---|---|---|
| Dustin Black | 2013-11-18 15:51:41 UTC | CC | dblack | |
| Brad Hubbard | 2013-11-18 20:45:18 UTC | Priority | unspecified | high |
| CC | bhubbard | |||
| Scott Haines | 2013-11-18 21:29:40 UTC | Keywords | ZStream | |
| CC | shaines | |||
| Nagaprasad Sathyanarayana | 2013-11-19 04:38:41 UTC | Keywords | ZStream | |
| CC | nsathyan | |||
| Anjana Suparna Sriram | 2013-11-19 10:38:42 UTC | CC | asriram | |
| Doc Text | Cause: h Consequence: Workaround (if any): Result: | |||
| Doc Type | Bug Fix | Known Issue | ||
| Pavithra | 2013-11-19 11:01:06 UTC | CC | psriniva | |
| Doc Text | Cause: h Consequence: Workaround (if any): Result: | Cause: h g Consequence: Workaround (if any): Result: |
||
| Pavithra | 2013-11-19 11:18:20 UTC | Doc Text | Cause: h g Consequence: Workaround (if any): Result: | During Geo-replication session, if the master node experiences a connectivity issue, the changes identified by the file system crawl process is not retained when the process is terminated. When the master nodes is back online, the file system crawl restarts and identifies the changes again. As a result, there is a performance degradation. |
| Vivek Agarwal | 2013-11-27 11:52:35 UTC | Keywords | ZStream | |
| Status | NEW | ASSIGNED | ||
| CC | vagarwal | |||
| Assignee | rhs-bugs | vshankar | ||
| Sudhir D | 2013-12-04 05:31:52 UTC | QA Contact | sdharane | vkoppad |
| Kotresh HR | 2013-12-11 08:17:56 UTC | CC | vshankar | |
| Venky Shankar | 2013-12-13 09:04:17 UTC | Assignee | vshankar | khiremat |
| Venky Shankar | 2013-12-13 09:06:28 UTC | Status | ASSIGNED | MODIFIED |
| Vivek Agarwal | 2013-12-17 07:35:39 UTC | Status | MODIFIED | ON_QA |
| Fixed In Version | glusterfs-3.4.0.50rhs | |||
| Harshavardhana | 2013-12-27 05:43:43 UTC | CC | fharshav | |
| Vijaykumar Koppad | 2014-01-02 06:37:04 UTC | QA Contact | vkoppad | vbhat |
| Kotresh HR | 2014-01-02 10:30:43 UTC | Doc Text | During Geo-replication session, if the master node experiences a connectivity issue, the changes identified by the file system crawl process is not retained when the process is terminated. When the master nodes is back online, the file system crawl restarts and identifies the changes again. As a result, there is a performance degradation. | Cause: The xtime updation was maintained only at root directory. It was not maintained per directory. Consequence: While doing first xsync crawl, disconnection with slave causes the geo-rep to re-crawl the whole file system and generate XSYNC-CHANGELOGS again. Fix: The fix is given to maintain the xtime per directory and is updated with slave's stime after the successful sync of all the files and subdirectories under it. This provides the infrastructure for the xsync to skip the directories which are already synced when it restarts. Result: xsync does not generate XSYNC-CHANGELOGS for already synced directory when it re-crawls the file system due to disconnection of the geo-rep session with the slave. |
| Doc Type | Known Issue | Bug Fix | ||
| M S Vishwanath Bhat | 2014-01-03 06:53:22 UTC | QA Contact | vbhat | vkoppad |
| Vijaykumar Koppad | 2014-01-03 13:36:11 UTC | Status | ON_QA | VERIFIED |
| QA Contact | vkoppad | vbhat | ||
| Vijaykumar Koppad | 2014-01-03 13:36:31 UTC | QA Contact | vbhat | vkoppad |
| Pavithra | 2014-01-08 10:03:49 UTC | CC | khiremat | |
| Doc Text | Cause: The xtime updation was maintained only at root directory. It was not maintained per directory. Consequence: While doing first xsync crawl, disconnection with slave causes the geo-rep to re-crawl the whole file system and generate XSYNC-CHANGELOGS again. Fix: The fix is given to maintain the xtime per directory and is updated with slave's stime after the successful sync of all the files and subdirectories under it. This provides the infrastructure for the xsync to skip the directories which are already synced when it restarts. Result: xsync does not generate XSYNC-CHANGELOGS for already synced directory when it re-crawls the file system due to disconnection of the geo-rep session with the slave. | Previously, when the first xsync crawl was in progress, disconnection with the slave volume caused Geo-replication to re-crawl the entire file system and generate XSYNC-CHANGELOGS. With this update, xsync skips the directories which are already synced to the slave volume. | ||
| Flags | needinfo?(khiremat) | |||
| Kotresh HR | 2014-01-08 10:58:17 UTC | Flags | needinfo?(khiremat) | needinfo+ |
| Jay Turner | 2014-01-15 11:25:54 UTC | Target Milestone | --- | RHS 2.1.2 |
| errata-xmlrpc | 2014-02-25 05:35:51 UTC | Status | VERIFIED | RELEASE_PENDING |
| errata-xmlrpc | 2014-02-25 08:04:24 UTC | Status | RELEASE_PENDING | CLOSED |
| Resolution | --- | ERRATA | ||
| Last Closed | 2014-02-25 03:04:24 UTC | |||
| John Skeoch | 2014-08-25 00:50:31 UTC | CC | bbandari | |
| Rejy M Cyriac | 2015-05-13 16:27:43 UTC | Target Release | --- | RHGS 2.1.2 |
| Target Milestone | RHS 2.1.2 | --- | ||
| Scott Haines | 2015-05-15 18:35:12 UTC | CC | shaines |
Back to bug 1031687