Re: maxlen in onstat -g ioq

Rüdiger, As I already pointed out, this is normal. It only shows that you do a substantial amount of writes at checkpoint time. What you could do is check your checkpoint duration times in the message log. If they are long and you don't like it because it blocks writers during that time, you can decrease LRU_MAX_DIRTY and LRU_MIN_DIRTY to let the flushers write more pages to disk between checkpoints and monitor it using "onstat -F". I doubt that increasing the number of LRUs will help. Michael Rüdiger Mähl wrote: > "Jimber Rio, Manuel" <manuel.jimber.sspa@juntadeandalucia.es> wrote: > > >>Are you system problems in performance? > > > Yes, we have slight performance problems, especially > compared to our test system, which run very well. The > upper class production system runs only a bit better. > > >>My system has maxlen very large, but the system has good response. You >>can change it adding more queues. > > > Ok, I will try it. > > Thanks a lot, > > Rüdiger -- === Michael Mueller ================== Tel. + 49 8171 63600 Fax. + 49 8171 63615 Web: http://www.mm.kay-mueller.de http://www.planets.kay-mueller.de ======================================

Michael Mueller <michael.mueller01@kay-mueller.de> wrote: > R'diger, > > As I already pointed out, this is normal. It only shows that you do a > substantial amount of writes at checkpoint time. What you could do is > check your checkpoint duration times in the message log. If they are > long and you don't like it because it blocks writers during that time, > you can decrease LRU_MAX_DIRTY and LRU_MIN_DIRTY to let the flushers > write more pages to disk between checkpoints and monitor it using > "onstat -F". Checkpoint duration is okay, but I will nevertheless change LRU_MIN/MAX_DIRTY. Thanks a lot, R'diger