Unsolved
This post is more than 5 years old
29 Posts
0
2360
Long running multi-scan job
Hi,
I have an issue with multi-scan running for a long time ( 5 days). We have not added any new nodes or anything to trigger an autobalance.
Does anyone else have experience with multiscan taking days to complete?
We are on 7.0.2.4.
Thanks
Peter_Sero
1.2K Posts
1
September 13th, 2014 02:00
Depends heavily on the cluster type, job impact setting, workload, avg file sizes.
Multiscan finishing with minutes -- I've only seen that on the virtual nodes so far...
Days or even weeks are quite commonly observed.
isi job status -v
gives some info on the progress (though the estimated finish time is not too reliable)
Multiscan always optimizes the whole file system,
independent of the initial status. So even if the system
appears well-balanced prior start, the job shuffles all the
bits around.
If it's unclear why the job started, I'd check with support.
-- Peter
cadiletta
106 Posts
0
September 15th, 2014 09:00
Has the multiscan made progress? Peter is spot on that there are a large number of factors to consider. But one question I'd ask is what impact is the job having on your workflow? Hopefully the job is running without significant degradation in performance, in which case it's just running low level in the background and you should be able to just allow it to slowly work through the cluster.
However if there are errors, or you are seeing performance problems or have other questions, I would discuss the issue with support, but let us know how things progress!