| Title | TeamTrack server fails when memory is low. |
| Status | closed |
| Priority | optional |
| Assigned user | Gareth Rees |
| Organization | TeamShare |
| Description | When the P4DTI fetches a lot of data from the TeamTrack server, eventually the P4DTI pauses for several minutes, and then reports an error from the TeamShare API. The error is: SERVER ERROR: Failure to receive return code: -1 |
| Analysis | Our recommended configuration is to run the P4DTI on the same machine as the TeamTrack server. This means that when the client reads a lot of data (and this is easy to do because the TeamShare API can take over 600k to represent a single case; see job000277) the machine runs out of memory and this affects the server before it affects the client. Analysis in a debugger shows that when the P4DTI hangs the TeamShare API client is waiting in a call to recv. Eventually there is a network timeout and the error can be printed. It's definitely not a memory error going unnoticed in the TeamShare API client: I tested running out of memory in the client, and Python catches that successfully and generates a MemoryError exception. So this is a TeamTrack server problem, not a TeamShare API client problem. Perhaps we should change our advice about the location of the P4DTI? This problem will be reduced in severity when we fix the capacity problem (job000277) because then memory won't be so easily exhausted. GDR 2001-07-13: Since job000277 was fixed, we haven't seen this problem. So I've closed the job. |
| How found | manual_test |
| Evidence | None. |
| Observed in | 1.1.1 |
| Created by | Gareth Rees |
| Created on | 2001-05-17 15:09:28 |
| Last modified by | Gareth Rees |
| Last modified on | 2001-12-10 19:43:34 |
| History | 2001-05-17 GDR Created. 2001-07-13 GDR Closed (see analysis). |
| Change | Effect | Date | User | Description |
|---|---|---|---|---|
| 12701 | closed | 2001-05-19 12:02:21 | Gareth Rees | Merged work from branch/2001-05-15/capacity to version/1.1. |