Bug #8290
Correct p2p queue limits for all pools
0%
Description
The p2p max limit settings on the newly added pool servers (50) differ from those on the rest of the pools (75).
This is currently only matters for the times when the data are replicated to other pools.
75 is considered a bit too high.
History
#1 Updated by Gerard Bernabeu Altayo over 5 years ago
Hi,
I've documented the process here:
https://cmsweb.fnal.gov/bin/view/Storage/DCache22Procedures#Bulk_change_dCache_pool_configuration
I'm fixing the different movers, will track the progress and exact commands in the ticket.
BTW: I can not edit the status of this tickets (or I don't know how)
Gerard
#2 Updated by Gerard Bernabeu Altayo over 5 years ago
ssh root@cmsdcacheadmindisk.fnal.gov
tmpfile=`mktemp`
for pool in `curl http://localhost:2288/queueInfo | grep cell | grep -E '(disk|tape)-disk' | sed 's/<td class="cell">//g' | sed 's/<\/td>//g'`; do echo -e "cd $pool\nmover set max active 1000\nmover set max active 1000 -queue=default\nmover set max active 1000 -queue=WAN\np2p set max active 50\npp set max active 50\nsave\n..\n"; done > $tmpfile
ssh -1 -i /etc/dcache/admin/server_key -c blowfish -p 22223 localhost -l admin < $tmpfile
#3 Updated by Gerard Bernabeu Altayo over 5 years ago
The previous commands fixed all inconsistencies & new pools will come with the proper movers (fix came from some puppet changes weeks ago).
Please resolve/close this ticket (as I can not).
#4 Updated by Natalia Ratnikova over 5 years ago
- Assignee set to Gerard Bernabeu Altayo
Hi Gerard,
Thanks, the limits for all nodes look consistent now.
I have changed the project settings in the redmine so that all members of the parent projects (Storage and DCSO) members have the same ability in cms-dcache.
I leave this task open, please try that you can close it now!
I now see that all members of DCSO now appear as co-assignees, not sure that we really want that.
#5 Updated by Gerard Bernabeu Altayo over 5 years ago
- Status changed from New to Resolved
Now I can edit :)