-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi,
Am 02.10.2014 um 12:08 schrieb William Hay: > One of the users on our cluster has submitted a series of jobs each dependent > on the previous job. The current lead job in this series is in the hqw state > although the job on which it depends has finished successfully according to > qacct. I am concerned that the -hold_jid dependency may not have been > correctly released despite the job on which it depends having finished. Is > there any way to determine why a job is in hqw state? If the dependency hold > has indeed not been released is there a way to manually release it? You can check the output `qstat -j <job_id>` of the waiting job. There you will find lines like: $ qstat -j 11728 ... jid_predecessor_list (req): test.sh jid_predecessor_list: 11726,11727 and the "jid_predecessor_list" will shrink for each completed job. - -- Reuti -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.20 (Darwin) Comment: GPGTools - http://gpgtools.org iEYEARECAAYFAlQ8AxMACgkQo/GbGkBRnRrxdwCfX7eba+hRA/oBe5ldeC83khce qroAnjzubdF3Tuebgxpw7pXyNf1kjPeM =NBDN -----END PGP SIGNATURE----- _______________________________________________ users mailing list users@gridengine.org https://gridengine.org/mailman/listinfo/users
