Last modified: 2014-04-24 16:06:45 UTC
Not sure if affects other migrated scripts as well, but for rsynced geowiki log files, group ownership changed from the stats group to the gmetric group. Old files (that have been owned by the stats group on stat1) are now owned by the group gmetric [1] on stat1003, while new files on stat1003 are again correctly owned by the stats group [2]. Did the gid of the stats group change between stat1 and stat1003? [1] ___________________________________________________________ stats@stat1003 // 0 // 09:11:58 cwd: /a/geowiki ll /a/geowiki/logs/process_data.py-cron-2014-04-02--12-00-01.stdout -rw-rw-r-- 1 stats gmetric 30660 Apr 2 12:32 /a/geowiki/logs/process_data.py-cron-2014-04-02--12-00-01.stdout [2] ___________________________________________________________ stats@stat1003 // 0 // 09:12:00 cwd: /a/geowiki ll /a/geowiki/logs/process_data.py-cron-2014-04-03--12-00-01.stdout -rw-rw-r-- 1 stats stats 32184 Apr 3 12:12 /a/geowiki/logs/process_data.py-cron-2014-04-03--12-00-01.stdout
I chgrp-ed in /a/geowiki. So /a/geowiki should be fixed. But as it seems to affect other parts as well [1], I am leaving the bug open so others (e.g.: ottomata, or whoever has more clue about those files in /a on stat1003) can chime in about them. [1] ___________________________________________________________ stats@stat1003 // 0 // 09:50:07 cwd: /a/geowiki ls -ld /a/reportcard-data /a/limn-public-data/mobile/datafiles/menu-daily.csv -rw-rw-r-- 1 stats gmetric 6232 Apr 22 09:04 /a/limn-public-data/mobile/datafiles/menu-daily.csv drwxr-xr-x 8 stats gmetric 4096 Jun 21 2012 /a/reportcard-data/
Hm, weird. Yeah probably some rsync + gid discrepancy. Fixed! find . -group gmetric -exec chgrp -v stats {} \;
Looks good to me. Thanks!