Back to bug 1378123
| Who | When | What | Removed | Added |
|---|---|---|---|---|
| Hynek Mlnarik | 2016-09-21 14:46:21 UTC | Status | NEW | POST |
| Hynek Mlnarik | 2016-09-23 13:54:49 UTC | Status | POST | MODIFIED |
| Fixed In Version | openstack-neutron-8.1.2-6.el7ost | |||
| Hynek Mlnarik | 2016-09-26 07:02:51 UTC | Status | MODIFIED | NEW |
| Doc Text | Cause: Due to concurrency, db error by precommit is sometimes inevitable. precommit may issue db transactions depending on mechanism driver, Some operations on e.g. subnet, routers, touches many other resources (e.g. ip allocation, ports) in single db transaction as well. And background operation by mechanism driver (e.g. background sync, monitoring resources periodically, etc) may touch those db tables with other order. Consequence: Transaction sometimes fails in precommit phase. Fix: Retry on fail. Result: Mitigation of the precommit issues. | |||
| Doc Type | If docs needed, set a value | Bug Fix | ||
| Hynek Mlnarik | 2016-09-26 10:03:44 UTC | Status | NEW | MODIFIED |
| errata-xmlrpc | 2016-11-21 18:15:04 UTC | Status | MODIFIED | ON_QA |
| Scott Lewis | 2016-11-22 20:54:18 UTC | Keywords | Triaged | |
| Toni Freger | 2016-11-23 10:45:06 UTC | QA Contact | tfreger | gcheresh |
| Charles Crouch | 2016-11-23 15:53:21 UTC | CC | charcrou | |
| GenadiC | 2016-11-24 10:06:53 UTC | Flags | needinfo?(hmlnarik) | |
| Hynek Mlnarik | 2016-11-24 13:16:58 UTC | Flags | needinfo?(hmlnarik) | needinfo?(tfreger) |
| GenadiC | 2016-11-24 13:44:56 UTC | Status | ON_QA | VERIFIED |
| Ofer Blaut | 2016-11-27 15:38:53 UTC | CC | oblaut | |
| Flags | needinfo?(tfreger) | |||
| Martin Lopes | 2016-11-30 03:24:19 UTC | CC | mlopes | |
| Doc Text | Cause: Due to concurrency, db error by precommit is sometimes inevitable. precommit may issue db transactions depending on mechanism driver, Some operations on e.g. subnet, routers, touches many other resources (e.g. ip allocation, ports) in single db transaction as well. And background operation by mechanism driver (e.g. background sync, monitoring resources periodically, etc) may touch those db tables with other order. Consequence: Transaction sometimes fails in precommit phase. Fix: Retry on fail. Result: Mitigation of the precommit issues. | Database precommit errors can occasionally occur due to concurrency. The precommit may issue database transactions, depending on the mechanism driver. Certain operations (for example, on subnets, routers) touch many other resources in a single database transaction (for example, IP allocation, ports). In addition, a background operation by the mechanism driver, (such as background sync, periodic resource monitoring) may touch those database tables in a different order. Consequently, a transaction may sometimes fail in the precommit phase. With this update, a retry is performed on the precommit failure. As a result, this can mitigate precommit issues. |
||
| Martin Lopes | 2016-11-30 03:34:50 UTC | Doc Text | Database precommit errors can occasionally occur due to concurrency. The precommit may issue database transactions, depending on the mechanism driver. Certain operations (for example, on subnets, routers) touch many other resources in a single database transaction (for example, IP allocation, ports). In addition, a background operation by the mechanism driver, (such as background sync, periodic resource monitoring) may touch those database tables in a different order. Consequently, a transaction may sometimes fail in the precommit phase. With this update, a retry is performed on the precommit failure. As a result, this can mitigate precommit issues. | Due to the nature of concurrency, database precommit errors can occasionally occur. The precommit may issue database transactions, depending on the mechanism driver. Certain operations (for example, on subnets, routers) touch many other resources in a single database transaction (for example, IP allocation, ports). In addition, a background operation by the mechanism driver, (such as background sync, periodic resource monitoring) may touch those database tables in a different order. Consequently, a transaction may sometimes fail in the precommit phase. With this update, a retry is performed on the precommit failure. As a result, this can mitigate precommit issues. |
| errata-xmlrpc | 2016-12-21 16:12:13 UTC | Status | VERIFIED | RELEASE_PENDING |
| errata-xmlrpc | 2016-12-21 16:48:35 UTC | Status | RELEASE_PENDING | CLOSED |
| Resolution | --- | ERRATA | ||
| Last Closed | 2016-12-21 11:48:35 UTC |
Back to bug 1378123