|cemc||got it in the meantime|
but thanks anyway
|mlo||Because I'd really know what they are talking about. Comparing a Network FS against a local FS is pretty dumb in the first place and I hoped a storage vendor would do better.|
cemc: Remember tho, last I checked RHat didnt offer idemnification if they later choose to reduce the advertised life span...
|cemc||mlo: i see|
|peerce||yah, that was my comment.|
|dan__t||Anyone ever seen an instance where syslog just kindof... stops? In order to get it to work again, it needs to be restarted. I've had this happen on at least half a dozen machines over the past, I'd say, year|
Usually I'd just ignore it or whatever, but it would be neat to see if there might be a solution for this.
|peerce||haven't seen that on any of my servers|
|cannonball||dan__t: Usually happens when logfiles are rotated out from under syslog without a -HUP being sent to the syslog daemon.|
|mlo||dan: Yes, but I dont know if its applicable to CentOS. Debian used to have syslogd compiled without LFS, so maximum log file size was 2.4Gig|
|dan__t||Ah, that would make sense.|
|peerce||yeah, are you using a non standard logrotate daemon?|
|dan__t||WEll, log file size isn't a problem, I think that cannonball's suggestion is more plausable.|
Nope, still using whatever comes with CentOS.
Otherwise I wouldn't have been able to ask. You know that ;)
cannonball, so this would be an issue with logrotate in regards to the wrong (or no) signal being sent?
|cannonball||Yes. Let me explain.|
|dan__t||I suppose syslog would keep logging to the last opened file handle that it had, which would be messages.1, .2, .3, etc etc.|
I understand how it works, but not too clear on why logrotate is not doing its thing properly.
|cannonball||Personally, I like to 'killall -HUP processname' when I can. But most scripts seem to want to 'kill -HUP `cat /var/run/processname.pid`' which means that if that processname.pid file is missing...|
ALright, well, that's something to look a.
|jsharper||for those talking about 3ware .. i see indications that a backported 3w-9xxx module that supports the 9650se may be in the 4.5 kernel|
unfortunately, that backport didnt make it into the 5.0 kernel
|KrnlKlink||anyone know if the areca drivers made it into 5.0?|
|irens||i don't know, sorry.|
|so__||KrnlKlink, if they're in RHEL5, they'll be in Centos5|
|irens||and CentOS 5 *maybe* will relase today, so stay tuned !|
|guma||jsharper: that is cool|
I have 3ware
I still have problem with 4.4 latest kernel update vs yum and Volume00 not found
Anyone has this problem?
|cannonball||No, I've never let the installer install an LVM system like that.|
|guma||cannonball: So you don't use LVM?|
|so__||I use lvm and I haven't had that problem|
|fozzmoo||pony when centos 5|