Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Set Java system property java.io.tmpdir to $VESPA_HOME/tmp | Harald Musum | 2020-03-12 | 1 | -0/+1 |
| | |||||
* | Use correct name of variable. | Henning Baldersheim | 2020-02-20 | 1 | -1/+1 |
| | |||||
* | Revert "Revert "Control -XX:CompressedClassSpaceSize jvm option."" | Henning Baldersheim | 2020-02-20 | 1 | -3/+8 |
| | |||||
* | Revert "Control -XX:CompressedClassSpaceSize jvm option." | Bjørn Christian Seime | 2020-02-20 | 1 | -8/+3 |
| | |||||
* | Control -XX:CompressedClassSpaceSize jvm option. | Henning Baldersheim | 2020-02-20 | 1 | -3/+8 |
| | |||||
* | Configure ZooKeeper log file handler | Harald Musum | 2020-01-13 | 1 | -3/+3 |
| | |||||
* | Set error file path | Valerij Fredriksen | 2020-01-09 | 1 | -0/+1 |
| | |||||
* | Stop requiring at least 1 Gb heap when using heaps size as percentage of ↵ | Harald Musum | 2019-12-18 | 1 | -4/+1 |
| | | | | | | | physical memory Make it easier to get this working for e.g. Docker containers with less memory | ||||
* | Log a message about the misconfiguration. | Henning Baldersheim | 2019-10-15 | 1 | -0/+1 |
| | |||||
* | Add a safety against specifying a min heap size lager than max heap size. | Henning Baldersheim | 2019-10-15 | 1 | -0/+5 |
| | |||||
* | minheapsize -> minHeapsize | Henning Baldersheim | 2019-10-15 | 1 | -3/+3 |
| | |||||
* | Add minheapsize as a controllable entity. Control it explicit and avoid ↵ | Henning Baldersheim | 2019-10-14 | 1 | -1/+3 |
| | | | | using append of jvmargs which is intended for user specified overrides. | ||||
* | VESPA_TOTAL_MEMORY_MB is not in use anymore | Harald Musum | 2019-10-11 | 1 | -9/+5 |
| | |||||
* | Add ActiveProcessorCount | Henning Baldersheim | 2019-08-20 | 1 | -1/+1 |
| | |||||
* | Since Java 11 and docker are note the best of friends we will use nproc instead. | Henning Baldersheim | 2019-08-20 | 1 | -0/+2 |
| | |||||
* | Use hash instead | Valerij Fredriksen | 2019-06-24 | 1 | -1/+1 |
| | |||||
* | Get memory limit from cgroup if possible and is set | Valerij Fredriksen | 2019-06-22 | 1 | -0/+5 |
| | |||||
* | Set -XX:ActiveProcessorCount option if set in config | Harald Musum | 2019-01-21 | 1 | -0/+9 |
| | | | | Use VESPA_CONTAINER_JVMARGS if set | ||||
* | Open java.base/java.io | gjoranv | 2019-01-21 | 1 | -0/+1 |
| | | | | | .. to allow com.yahoo.io.NativeIO to use reflective access on java.io.FileDescriptor.fd | ||||
* | nobody uses JSVC aka commons.daemon, remove support for it | Arne Juul | 2019-01-21 | 1 | -82/+0 |
| | |||||
* | CMS is deprecated in JDK 9 and will be removed in a later release | gjoranv | 2019-01-21 | 1 | -1/+1 |
| | | | | - See http://openjdk.java.net/jeps/291 | ||||
* | Open necessary packages to the unnamed module | gjoranv | 2019-01-21 | 1 | -0/+8 |
| | | | | | | .. to avoid warnings due to Felix accessing non-public apis with reflection. + Show all "Illegal reflective access" warnings. | ||||
* | Let the fallback be the old default. | Henning Baldersheim | 2018-10-25 | 1 | -1/+1 |
| | |||||
* | Setting NewRatio is not recommended for G1 | Bjørn Christian Seime | 2018-10-24 | 1 | -1/+1 |
| | | | | | | From G1 doc: "Avoid explicitly setting young generation size with the -Xmn option or any or other related option such as -XX:NewRatio. Fixing the size of the young generation overrides the target pause-time goal." | ||||
* | Use G1GC as default | Bjørn Christian Seime | 2018-10-24 | 1 | -1/+1 |
| | |||||
* | Revert "Use G1GC as default" | Bjørn Christian Seime | 2018-10-20 | 1 | -1/+1 |
| | |||||
* | Setting NewRatio is not recommended for G1 | Bjørn Christian Seime | 2018-10-19 | 1 | -1/+1 |
| | | | | | | From G1 doc: "Avoid explicitly setting young generation size with the -Xmn option or any or other related option such as -XX:NewRatio. Fixing the size of the young generation overrides the target pause-time goal." | ||||
* | Use G1GC as default | Bjørn Christian Seime | 2018-10-19 | 1 | -1/+1 |
| | |||||
* | Revert "Revert "Explicitly use legacy format option to vespa-get-config"" | Harald Musum | 2018-09-04 | 1 | -1/+1 |
| | |||||
* | Revert "Explicitly use legacy format option to vespa-get-config" | Arnstein Ressem | 2018-08-29 | 1 | -1/+1 |
| | |||||
* | Explicitly use legacy format option to vespa-get-config | Harald Musum | 2018-08-29 | 1 | -1/+1 |
| | |||||
* | just exit on out of memory | Arne Juul | 2018-07-18 | 1 | -2/+2 |
| | | | | | | * asking java to fork 'kill -9 %p' will often fail with "os::fork_and_exec failed: Cannot allocate memory (12)" message. * instead, use a newish option (from JDK 8u92) to just exit | ||||
* | Set MaxJavaStackTraceDepth to a value legal for all JVM versions. | gjoranv | 2018-07-05 | 1 | -2/+2 |
| | | | | - On JDK 9+, -1 is invalid, and 0 really means zero. | ||||
* | Revert "Use RTMLocking" | Harald Musum | 2017-12-12 | 1 | -2/+0 |
| | |||||
* | Use RTMLocking | Henning Baldersheim | 2017-12-12 | 1 | -0/+2 |
| | |||||
* | The cost of preserving the framepointer is small enough to allow it being ↵ | Henning Baldersheim | 2017-09-27 | 1 | -0/+2 |
| | | | | | | | default. Performance information benefits is significant and will over time counter the cost. Observed cost has been below the noise level in the 1% range. | ||||
* | Use vespa-get-config | Harald Musum | 2017-08-16 | 1 | -1/+1 |
| | |||||
* | Use full path to getvespaconfig | Harald Musum | 2017-08-16 | 1 | -1/+1 |
| | |||||
* | VESPA_HOME should not end with / anymore | Arne H Juul | 2017-06-17 | 1 | -16/+16 |
| | | | | | * best practice for HOME-like env vars is to be just the name of the directory without trailing slash. | ||||
* | Update copyright headers | Jon Bratseth | 2017-06-14 | 1 | -1/+1 |
| | |||||
* | Revert "Update copyright headers" | Jon Bratseth | 2017-06-14 | 1 | -1/+1 |
| | |||||
* | Update copyright headers | Jon Bratseth | 2017-06-14 | 1 | -1/+1 |
| | |||||
* | Revert "Copyright header" | Jon Bratseth | 2017-06-13 | 1 | -1/+1 |
| | |||||
* | Copyright header | Jon Bratseth | 2017-06-13 | 1 | -1/+1 |
| | |||||
* | Revert "Run extension script if exists" | Bjørn Christian Seime | 2017-05-15 | 1 | -8/+0 |
| | | | | This reverts commit 1e7efd7ddabb9fff21c9d76a2ff764bd7b740cd1. | ||||
* | Run extension script if exists | Bjørn Christian Seime | 2017-05-08 | 1 | -0/+8 |
| | |||||
* | Also change start container daemon | Håkon Hallingstad | 2017-04-05 | 1 | -2/+2 |
| | |||||
* | Only update jvm_heapsize if percentage has been set | Håkon Hallingstad | 2016-10-05 | 1 | -16/+14 |
| | |||||
* | Use jvm_heapsize for container's Java process if explicitly set | Håkon Hallingstad | 2016-10-03 | 1 | -16/+15 |
| | | | | | | | | | Also simplifies the logic a bit: - If jvm_heapsize has been set by model, use that - Otherwise, if jvm_heapSizeAsPercentageOfPhysicalMemory has been set, use that to calculate the heapsize. Takes into account that TOTAL_MEMORY_MB may have been set to override the size of the physical memory. - Otherwise, use 1536MB. | ||||
* | Handle absent heapSizeAsPercentageOfPhysicalMemory | Håkon Hallingstad | 2016-09-28 | 1 | -2/+7 |
| |