vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| "Casper H.S. Dik" <Casper.Dik@Sun.COM> schrieb im Newsbeitrag news:4017c49b$0$332$e4fe514c@news.xs4all.nl... > When you install patches with a logging "/" you should > "lockfs -f /" prior to doing a reboot. > > Until the code is written which allows booting with a dirty log, > it can't be the default. Speaking of installing patches (perhaps this should be made into another thread), would you know WHY the latest master patch cluster for Solaris7 causes the kernel to panic in an endless loop and makes the system highly unstable on a SUNW,Sun_4_50? Kernel patch 106541-31. Just please don't tell me "Sun doesn't support Solaris7 on Sun_4_50 any more..." |
| |||
| On Fri, 30 Jan 2004, UNIX admin wrote: > "David Combs" <dkcombs@panix.com> schrieb im Newsbeitrag > news:bv97jp$gue$1@reader2.panix.com... > > > (Before I go to google) What's "fastfs"? No manpage in s9. > > > > Is fastfs something that *lots* (even most of the "knowledgeable" (sp?) > > solaris admins) use? Or at least know about?) > > He he he, no, no, what he means is that the FileSystem is supposed to > perform faster when logging is off. No, fastfs is a little utility that can speed up a UFS file system. > One way or another, turn logging on on all your FileSystems except swap. Good advice, although swap is not a file system... -- Rich Teer, SCNA, SCSA President, Rite Online Inc. Voice: +1 (250) 979-1638 URL: http://www.rite-online.net |
| |||
| jayl-news@accelerant.net (Jay Lessert) wrote in message news:<7109f92b.0401291245.2bdc0d96@posting.google. com>... > ultrasparc3@hotmail.com wrote in message news:<2004028.204651.3924@cable.prodigy.com>... > > are you aware that the boot code only understands plain (non-logging) > > UFS? if any files necessary early in the boot process have their metadata > > in the log then you will have problems. > > You mean if I yank the power plug right after installing a kernel > jumbo patch? But I'm probably due for problems in that case, > logging or not. Typically, I've run into it if a kernel panic (for me, mostly from reboot -d) occurred right after adding a driver, or too quickly after a reconfiguration boot. Typically, either /etc/path_to_inst or /etc/name_to_major has inconsistent metadata, causing the kernel to become quite confused, and keeping boot from completing. This happens much more often for kernel developers than it does for most customers, but the subtlety of the errors which can occur means that the current plan is to only turn on logging by default when ufsboot supports using it. - jonathan |
| |||
| In article <oM2Sb.3790$2J1.1801@fe1.texas.rr.com>, Logan Shaw <lshaw-usenet@austin.rr.com> wrote: >David Combs wrote: >> (Before I go to google) What's "fastfs"? No manpage in s9. >> >> Is fastfs something that *lots* (even most of the "knowledgeable" (sp?) >> solaris admins) use? Or at least know about?) > >It's less useful these days, now that logging exists. >But it still sometimes can be useful. > >Here's the scoop: filesystems need to write certain data in Thanks so much for the super explanation! (Why doesn't SUN (with permission from posters like you) take some of this stuff and put it into a large book or something, and let us all read it?) David |
| |||
| On Sat, 31 Jan 2004, David Combs wrote: > (Why doesn't SUN (with permission from posters like you) > take some of this stuff and put it into a large book or > something, and let us all read it?) They do; it's called groups.google.com... :-) -- Rich Teer, SCNA, SCSA President, Rite Online Inc. Voice: +1 (250) 979-1638 URL: http://www.rite-online.net |
| |||
| Actually http://www.sun.com/bigadmin "Rich Teer" <rich.teer@rite-group.com> wrote in message news:Pine.SOL.4.58.0401311312070.13847@zaphod.rite-group.com... > On Sat, 31 Jan 2004, David Combs wrote: > > > (Why doesn't SUN (with permission from posters like you) > > take some of this stuff and put it into a large book or > > something, and let us all read it?) > > They do; it's called groups.google.com... :-) > > -- > Rich Teer, SCNA, SCSA > > President, > Rite Online Inc. > > Voice: +1 (250) 979-1638 > URL: http://www.rite-online.net |
| |||
| In article <sb_Sb.115365$cM1.20334051@twister.nyc.rr.com>, Rodrick Brown <rbrown@[remove]doitt.nyc.gov> wrote: >Actually http://www.sun.com/bigadmin Thanks --I'm on my way over there now! David |
| ||||
| ultrasparc3@hotmail.com wrote: > In <40174373@news.nucleus.com> cbigam@somewhereelse.nucleus.com writes: >>> Fourth -- if you do use it, do you turn it on for *all* >>> file-systems, or only some of them? (And why some and not others?) >> >>Every filesystem on every server, INCLUDING / and /var. Some have avoided >>putting it on /, but I have found no reason to do so whatsoever. > > are you aware that the boot code only understands plain (non-logging) > UFS? if any files necessary early in the boot process have their metadata > in the log then you will have problems. Indeed I am, and ran into this problem about a week before posting my blanket recommendation. The problem was that someone had plugged a keyboard into a headless (and very very remote) server, in the middle of many changes being written to / and /var. The result was a box that couldn't boot in single user mode. So I booted from CD, turned off logging, manually unmirrored the disks (not bad with SDS--just remember to comment out the lines in /etc/system!), fscked the drive with an alternate superblock, and booted without a problem. Reenabling logging was a bit annoying, but that was a minor issue since the box was alive and healthy by that point. So I stick by my statement, even in the worst case scenario. Colin |