Justin S. Leitgeb wrote: > Fyodor seems to be in pretty bad shape since the disk installation. I'm > not sure if these things are related, but I logged in again today to see a > 30 + load average and about 25 processes blocked according to vmstat. > This is why I was concerned that a completely different model and geometry drive was being used for mirroring. Of course it "works," but that doesn't mean it works well, or fast. > So, I think that if we're not ready to really start migrating in the next > few days, we have to do some performance tuning or troubleshooting. What > do y'all think? > Yes. > Second, there was a process called blosxom.cgi under terry's user account. > It's a perl blogging program. I noticed it was using a good bit of CPU also, which is odd since it's supposed to be fairly lightweight and terry's sites are not the busiest lately. Terry, could you take a look at your site's log or config to make sure it's working right? It may be contributing, but I don't think it's more than a small share of the problem. > Also, is there something that could be related to the new disk that is > causing these issues? > Since I'm not root on fyodor and dmesg is disabled, I don't know. Can someone post dmesg|grep hd to -sysadmin? Maybe they gave us a slow drive. > Sorry to distract people from the migration tasks, but I think that we > should probably address these issues that are now causing lots of > down-time every day recently. > I also think we should start migrating the busiest websites ASAP. We can start this weekend. I'm willing to move mine.
-Nathan _______________________________________________ HCoop-SysAdmin mailing list [email protected] http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-sysadmin
