[gpfsug-discuss] Can 5-minutes frequent lsscsi command disrupt GPFS I/O on a Lenovo system ?

[Dorigo Alvise (PSI)]

Hello folks,
recently I observed that calling every 5 minutes the command "lsscsi -g" on a Lenovo I/O node (a X3650 M5 connected to D3284 enclosures, part of a DSS-G220 system) can seriously compromise the GPFS I/O performance.

(The motivation of running lsscsi every 5 minutes is a bit out of topic, but I can explain on request).

What we observed is that there were several GPFS waiters telling that flushing caches to physical disk was impossible and they had to wait (possibly going in timeout).

Is this something expected and/or observed by someone else in this community ?

Thanks
Regards,

   Alvise Dorigo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20190916/479fd0fe/attachment-0001.htm>