|
Mantis - Resin
|
|||||
| Viewing Issue Advanced Details | |||||
|
|
|||||
| ID: | Category: | Severity: | Reproducibility: | Date Submitted: | Last Update: |
| 6291 | minor | always | 10-21-19 05:07 | 03-21-24 07:10 | |
|
|
|||||
| Reporter: | nam | Platform: | |||
| Assigned To: | OS: | ||||
| Priority: | high | OS Version: | |||
| Status: | new | Product Version: | 4.0.62 | ||
| Product Build: | Resolution: | open | |||
| Projection: | none | ||||
| ETA: | none | Fixed in Version: | |||
|
|
|||||
| Summary: | 0006291: mod_caucho does not see new elastic servers | ||||
| Description: |
(rep by J. Meyer) mod_caucho does not automatically pick up new elastic servers. The elastic servers should show up under the virtual host section. The issue appears to Resin's failure to update the in-memory host etag cache after a cluster change. The workarounds are either: 1. restarting the triad servers, thereby clearing the cache on the backend, or 2. restarting apache/mod_caucho, thereby clearing the cache on the frontend If only a triad server is restarted, then mod_caucho would eventually see the elastic server when it queries the restarted triad server through round-robin. code trail: 1. HmuxDispatchRequest.handleRequest() 2. HmuxDispatchRequest.queryAll() 3. Host.setConfigETag() An alternative workaround is to prevent mod_caucho from sending the etag to Resin. In modules/c/src/common/config.c->cse_update_host_from_resin(), delete the following lines: if (host->etag[0]) hmux_write_string(&s, HMUX_DISPATCH_ETAG, host->etag); |
||||
| Steps To Reproduce: | |||||
| Additional Information: | |||||
| Relationships | |||||
| Attached Files: |
|
||||
| There are no notes attached to this issue. |