We have found an issue in HTCondorÂversion 8.8.1 regarding condor_annex command. We are trying to test some resources in AWS to extend our HTCondor Cluster. We saw that from the last stable version 8.8.1 the command is available, so it came at the right time.
The problem is that we got a Segmentation Fault error, and we would like to tell you what we have done so you can track the problem.
We thought we were doing somethingÂstrangeÂbecause we had to put this in place with our existing Condor configuration, so we reproduced the condor_annex in a test instance to create a "personal" condor environment, but the error appeared again. Then we decided to test with the 8.9.0 development version, and we were able to make it work.
Using version 8.8.1, after running the first steps successfully, the segmentation fault appears when we tried to create the annex:
[jcasals@cmxx .condor]$ condor_annex -count 2 -annex-name pic_annex
condor_annex -count 1 -annex-name pic_annex
Will request 1 m4.large on-demand instance for 0.83 hours.Â Each instance will terminate after being idle for 0.25 hours.
Is that OK? Â(Type 'yes' or 'no'): yes
Segmentation fault (core dumped)
The AnnexLog where there is the stack dump show us the next lines:
03/29/19 15:47:08 GAHP server pid = 850285
Caught signal 11: si_code=2, si_pid=17687328, si_uid=0, si_addr=0x10DE320
Stack dump for process 850280 at timestamp 1553870828 (4 frames)
Thus, after seeing this, we updated the condor version toÂ8.9.0 and reproduce the same steps, then, there are no more segfault errors.
Thank you in advance.