DSS - dyadic snapshot scheduler https://git.tuebingen.mpg.de/cgit/dss/atom?h=maint 2017-11-13T16:21:57Z 2017-11-13T16:21:57Z Andre Noll maan@tuebingen.mpg.de 2017-11-11T05:16:51Z urn:sha1:404aae666e5f8d7ce511639cbe4727a7f2e690cc If the oldest snapshot has to be removed because disk space is low, we currently log "orphaned" as the reason, which is incorrect. 2017-04-16T11:43:20Z Andre Noll maan@tuebingen.mpg.de 2017-04-16T11:43:20Z urn:sha1:965855384d34b7f1c4730eb78f26d010adb9be42 A couple of fixes and improvements have accumulated over the last year, so here's dss-0.1.7. Most likely this is going to be the last 0.1.x release. 2017-04-16T11:25:10Z Andre Noll maan@tuebingen.mpg.de 2017-04-16T11:25:10Z urn:sha1:71d84d1ad6047d5d24cc6c74d7b1493e8a995f74 This avoids dss.c:1519:32: warning: declaration of 'params' shadows a previous local [-Wshadow] The warning is harmless, though. 2017-04-16T10:32:06Z Andre Noll maan@tuebingen.mpg.de 2017-02-17T14:36:50Z urn:sha1:ae647da612e0edfb607a4fdb251983d6405f6b00 This function is rather big, so it's not clear whether it should be inlined or not. Without the inline attribute, that's up to the compiler to decide. 2017-04-16T10:32:06Z Andre Noll maan@tuebingen.mpg.de 2017-02-17T14:15:10Z urn:sha1:b795331d8f7006e1cd15ddf124baf1384347d2db The function only reads from the location pointed at by the config_file variable. 2017-04-16T10:32:06Z Andre Noll maan@tuebingen.mpg.de 2017-04-16T10:18:11Z urn:sha1:01af36ee3cd794f4332c4898828a31aafe0a939f In the error case do_semop() already returns the error code which corresponds to errno. 2016-08-23T11:26:25Z Andre Noll maan@tuebingen.mpg.de 2016-08-23T11:26:25Z urn:sha1:ee132d4f3e81eaaaeee94ceaf9f960c50bc7d448 The log message that prints the name of the snapshot which is going to be recycled is kind of interesting and deserves the "notice" severity. The "creating new snapshot" message is a bit misleading in case we decided to recycle a snapshot, so drop the word "new". 2016-08-21T19:13:25Z Andre Noll maan@tuebingen.mpg.de 2016-06-17T07:18:40Z urn:sha1:03918564cf43cfbd4f873f1b9c6ec3ccd15ede91 There is no need to actually obtain the lock. A single semaphore operation will do just fine. With sem_op equal to zero and IPC_NOWAIT the semop() call returns immediately, and the return value tells whether the semaphore value was zero. Rename the (static) function to mutex_is_locked() to indicate that it performs only read-only operations on the semaphore set. 2016-08-21T19:13:25Z Andre Noll maan@tuebingen.mpg.de 2016-06-16T21:06:29Z urn:sha1:81d9c8b0e49a12f7b726b06e2df48c2c3314bbcb This changes get_dss_pid() to handle the case where the caller passed a NULL pid pointer. Conversely, if pid is not NULL, we now make sure to initialize the given address in all cases. The single caller currently never passes NULL, so this change is just defensive programming, protecting against future users. Be liberal in what you accept, be strict in what you return.. 2016-08-21T19:13:25Z Andre Noll maan@tuebingen.mpg.de 2016-06-14T08:40:22Z urn:sha1:7da9a331fbbfb10a35fd9d014c47edbad13ffed7 This is pointless because we immediately cd back into the destination dir anyway.