Re: [OpenAFS] Summary of recommended configuration options from the workshop
In message [EMAIL PROTECTED],Esther Filderman writes: With Demand Attach, the file servers can make your coffee, do your laundry and change the baby's diaper. Demand Attach can end the wars in the Middle East, stop racism in it's tracks and reduce the effects of Global Warming. With Demand Attach, you'll never have your Windows boxes BSOD, your Unix boxes will never need to fsck, and you'll get a shiny US nickel, as well! no x-ray vision? nfs has x-ray vision. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On May 26, 2008, at 3:48 PM, Russ Allbery wrote: . . . Plus, a stable demand-attach is a good milestone for releasing 1.5 . . . Agreed. That said, do we have a milestone list for 1.5 becoming 1.6? ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
Steve Simmons [EMAIL PROTECTED] writes: On May 26, 2008, at 3:48 PM, Russ Allbery wrote: . . . Plus, a stable demand-attach is a good milestone for releasing 1.5 . . . Agreed. That said, do we have a milestone list for 1.5 becoming 1.6? At this point, I'm fairly sure that the only major thing that's left is testing and fixing the resulting bugs, although Derrick is the best person to give a canonical answer. -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/ ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On May 26, 2008, at 8:16 PM, Esther Filderman wrote: On Mon, May 26, 2008 at 6:43 PM, Robert Banz [EMAIL PROTECTED] wrote: however, with 1.4, fastrestart is teh bomb. ;) Yes, but it's teh bomb that can blow up in your face. Remember that the salvager is there to make sure you're volumes aren't falling apart. With fast-restart you can theoretically find your volumes suddenly falling out of service (or worse, serving up crap) because of corruption. I'd happily write a script/whathaveyou that looks over the unmounted volumes and one by one salvages them. It strikes me as the kind of things you'd want to start running 30 minutes or so after the file server restarts. A swatch-like script that scans for volumes that need salvage would be another win. Recent discussion here touted the utility of doing an occasional salvage as a best practice thing. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On May 29, 2008, at 2:16 PM, Russ Allbery wrote: Steve Simmons [EMAIL PROTECTED] writes: On May 26, 2008, at 3:48 PM, Russ Allbery wrote: . . . Plus, a stable demand-attach is a good milestone for releasing 1.5 . . . Agreed. That said, do we have a milestone list for 1.5 becoming 1.6? At this point, I'm fairly sure that the only major thing that's left is testing and fixing the resulting bugs, although Derrick is the best person to give a canonical answer. Yet another good reason for us to bring up a test server with it. We're in early deployment of this summer's upgrades; as soon as the initial set is up we're going to be trying out a test 1.5 server. Yes, Moose, this means you. :-) Steve ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
Steve Simmons wrote: I'd happily write a script/whathaveyou that looks over the unmounted volumes and one by one salvages them. It strikes me as the kind of things you'd want to start running 30 minutes or so after the file server restarts. A swatch-like script that scans for volumes that need salvage would be another win. Recent discussion here touted the utility of doing an occasional salvage as a best practice thing. Regardless, Demand Attach is the best answer we have available. With DAFS you can restart your file servers and the clients never have to be sent callback notifications because all of the existing callbacks are restored after the restart. With Demand Attach, the file servers can operate faster since fewer volumes are tracked in memory when not in use. smime.p7s Description: S/MIME Cryptographic Signature
Re: [OpenAFS] Summary of recommended configuration options from the workshop
Correct me if I'm wrong, but I seem to recall someone mentioning that there were certain cases when running fastrestart where volumes might end up being attached even if they need salvaging, leading to data loss/corruption? I would say any benefit you see in running fastrestart would be taken over by the chance that you could lose entire volumes to such a bug. I say the sooner we can get DAFS / 1.5 stable the better. DAFS should make your fileservers restart real fast too. -- Andy ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
Andy Cobaugh wrote: Correct me if I'm wrong, but I seem to recall someone mentioning that there were certain cases when running fastrestart where volumes might end up being attached even if they need salvaging, leading to data loss/corruption? Not corruption but there is a chance that the damage to the volume would not prevent it from being attached and the error would only be detected when a user actually attempts to access/change the contents. At that point the file server would notice the damage and detach the volume making it inaccessible to users. Jeffrey Altman smime.p7s Description: S/MIME Cryptographic Signature
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On Thu, May 29, 2008 at 2:25 PM, Steve Simmons [EMAIL PROTECTED] wrote: Yet another good reason for us to bring up a test server with it. We're in early deployment of this summer's upgrades; as soon as the initial set is up we're going to be trying out a test 1.5 server. Yes, Moose, this means you. :-) Oh, bite me. I'M the one who told YOU we should run this shit. :-P Moose ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On Thu, May 29, 2008 at 3:03 PM, Jeffrey Altman [EMAIL PROTECTED] wrote: Andy Cobaugh wrote: Correct me if I'm wrong, but I seem to recall someone mentioning that there were certain cases when running fastrestart where volumes might end up being attached even if they need salvaging, leading to data loss/corruption? Not corruption but there is a chance that the damage to the volume would not prevent it from being attached and the error would only be detected when a user actually attempts to access/change the contents. At that point the file server would notice the damage and detach the volume making it inaccessible to users. Well, it's theoretically possible for a damaged volume to get mounted, it's theoretically possible that it's in such a state that it tries to serve data that's been corrupted. It's theoretically possible that the cachemanager doesn't detect the problem and delivers crap to the user. It's theoretically possible that I'm going to wake up tomorrow less bitter, that people will stop misspelling Derrick's name, and that I'm going to win millions of dollars in the lottery. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On Thu, May 29, 2008 at 2:35 PM, Jeffrey Altman [EMAIL PROTECTED] wrote: Regardless, Demand Attach is the best answer we have available. With DAFS you can restart your file servers and the clients never have to be sent callback notifications because all of the existing callbacks are restored after the restart. With Demand Attach, the file servers can operate faster since fewer volumes are tracked in memory when not in use. With Demand Attach, the file servers can make your coffee, do your laundry and change the baby's diaper. Demand Attach can end the wars in the Middle East, stop racism in it's tracks and reduce the effects of Global Warming. With Demand Attach, you'll never have your Windows boxes BSOD, your Unix boxes will never need to fsck, and you'll get a shiny US nickel, as well! Get Demand Attach AFS NOW! [This offer only available in certain places. Offer valid only in countries whose names don't end in an alphanumeric character. Six box tops from DAFS required for each offer. Qualities of DAFS may be exaggerated for demonstration purposes. Not all cells will get the full special effects of DAFS. DAFS may require protective gear while in use, including but not limited to helmets, goggles, gloves, galoshes, steel-toed boots, non-reactive metal underwear, wool socks and/or leg warmers. The makers of DAFS disclaim any prior knowledge of your existence, and will deny it if you happen to be caught or killed. Demand Attach should never be used for medical purposes, including contraception. DAFS is not meant to be taken internally, nor topically, nor tropically. This offer void if seal is broken, but you can break the walrus all you want. This message emphatically not endorsed by my new owners^Wemployers, the University of The US State That Looks Like A Mitten.] ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
Re: [OpenAFS] Summary of recommended configuration options from the workshop
On Thu, May 29, 2008 at 7:17 PM, Esther Filderman [EMAIL PROTECTED] wrote: On Thu, May 29, 2008 at 2:25 PM, Steve Simmons [EMAIL PROTECTED] wrote: Yet another good reason for us to bring up a test server with it. We're in early deployment of this summer's upgrades; as soon as the initial set is up we're going to be trying out a test 1.5 server. Yes, Moose, this means you. :-) Oh, bite me. I'M the one who told YOU we should run this shit. :-P Whoops, that wasn't meant to be sent to the list. Sorry, folks. Gee, I'm, er, new to this whole newfangled email thing. Yeah, that's it. ___ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info