OK so looking back through the docs etc I have some what answered my own question. my rr_min_io values is pretty out of the ordinary<br>from reading through things. Also I could adjust the polling_interval back to the default of 5
<br><br>my multipath.conf default looks like (as mentioned before) this:<br clear="all">========= /etc/multipath.conf ================<br>defaults {<br> multipath_tool "/sbin/multipath -v0"<br> udev_dir /dev
<br> polling_interval 2<br> selector "round-robin 0"
<br> path_grouping_policy failover<br> getuid_callout "/sbin/scsi_id -g -u -s /block/%n"<br> path_checker readsector0<br> prio_callout "/bin/true"
<br> features "0"<br> rr_min_io 2 <br> rr_weight priorities<br> failback immediate<br>
no_path_retry fail<br>
user_friendly_name yes<br>}<br>========= ============= ================<br>
<br>I do not replace the rr_min_io value within my multipath device sections so a map looks like this when complete.<br><br> 0 33576960 multipath 0 0 2 1 round-robin 0 1 1 8:80 2 round-robin 0 1 1 8:176 2<br> ^
<br> |_ this value is way off from the default of 1000 <br><br>I know this is a highly user specific value however do most people stay around the default value for this setting.
<br>Specifically I'm looking for input on an io value for db volumes. <br><br>Also would all this be a non issue if I set rr_weight to uniform instead of priorities ( which I think evals to (0*2) in my case ).<br>Any help appreciated..
<br><br>-- <br>:wq!<br>kevin.foote