Just tested it again personally on my farm - works like a charm, slaves being attached with exact the same snapshot-based volume and mounted according to setting. I am seeing in your log that you're manually detaching and attaching Auto-EBS volumes, which is not a right thing to do.
On 28 дек, 09:45, Ken M <[email protected]> wrote: > Not sure if other EBS issues are resolved, but now I'm having sort of > the opposite problem. > > With the new script to prevent snapshot reloading on mysql machines > with ebs, the master reboots and attaches sucessfully to the last used > volume. > > However, the slaves are getting the most recent snapshots at all. > Even though the role is configured to auto-attach to snapshot, it > isn't for me on the slaves. > > On Dec 14, 4:28 am, Alex Kovalyov <[email protected]> wrote: > > > > I have just run across this problem as well. > > > This apparently wasn't a problem? - Kevin was detaching auto-attached > > volumes by hands. > > > > When I terminated the farm, I selected to keep theEBSvolumes, > > > however, when the farm terminated, the volumes completely disappeared > > > and had to be re-created on boot. This also happened when I did a > > > "synchronize to all" on the instance. This defeats the purpose of > > >EBS... and really screws up the ability to keep a persistent copy of > > > up-to-the-minute data, which makes it impossible to keep logfiles or > > > run applications that change their state (like users uploading images, > > > etc). I don't see anything in the logs regarding the volumes being > > > deleted. The farm id in question is 959 if that helps. > > > Do you know IDs of volumes that disappeared? > > Can you help us to reproduce this? > > Thanks alot. > > > On 12.12.08 01:23, "jmochs" <[email protected]> wrote: > > > > I have just run across this problem as well. > > > > When I terminated the farm, I selected to keep theEBSvolumes, > > > however, when the farm terminated, the volumes completely disappeared > > > and had to be re-created on boot. This also happened when I did a > > > "synchronize to all" on the instance. This defeats the purpose of > > >EBS... and really screws up the ability to keep a persistent copy of > > > up-to-the-minute data, which makes it impossible to keep logfiles or > > > run applications that change their state (like users uploading images, > > > etc). I don't see anything in the logs regarding the volumes being > > > deleted. The farm id in question is 959 if that helps. > > > > Thanks, > > > James > > > > On Dec 11, 2:56 pm, Kevin Baker <[email protected]> wrote: > > >> On Dec 11, 2008, at 3:53 PM, Kevin Baker <[email protected]> wrote: > > > >>> On Dec 10, 2008, at 9:52 AM, Alex Kovalyov <[email protected]> > > >>> wrote: > > > >>>> Dear Kevin, > > > >>>> We've spend few hours analyzing your logs and came to conclusion > > >>>> that you > > >>>> were detaching volumes that Scalr attaches for you automatically and > > >>>> attaching another one manually. > > >>>> Is this a right conclusion? > > > >>> Thank you tons for taking the time. Yes this the correct conclusion. > > >>> However I am trying to setup a confit very similiar to Ken M. > > > >>> I want myEBSTo automatically remount to an instance. The snapshot > > >>> is fine for backups, but the goal is to use the inherant data > > >>> persistance features of theEBS. This would require binding anEBS > > >>> to a role and having any new instances scalr boots to a role check > > >>> for orphanedEBSinstances. > > > >>> I was understanding > > > >> Oops, hut the send button ;) > > > >> It was my understanding. From the scalr role wizard that this was > > >> standard behavior. > > > >>>> If you want some particular volume to be attached, just make a > > >>>> snapshot of > > >>>> it, and choose " Attach volume from snapshot:" onEBStab. > > > >>>> On 10.12.08 17:45, "Kevin Baker" <[email protected]> wrote: > > > >>>>> I will be retesting with a fresh image today, and will update this > > >>>>> thread. > > > >>>>> On Dec 10, 2008, at 1:21 AM, Sam <[email protected]> wrote: > > > >>>>>> This is worrying as my application heavily rely on the > > >>>>>> application to > > >>>>>> attach to the already existingEBSand if it attaches newEBS > > >>>>>> everytime my application will fail. > > >>>>>> Kevin, did you recently checked this after a few problems were > > >>>>>> fixed > > >>>>>> related toEBS? > > > >>>>>> On Dec 9, 7:28 pm, Kevin Baker <[email protected]> wrote: > > >>>>>>> Alex, > > > >>>>>>> I have had anEBSinstance attach successfully each time, however > > >>>>>>> it is > > >>>>>>> always a newEBSinstance. > > > >>>>>>> It goes something like this: > > > >>>>>>> * CreateEBS > > >>>>>>> * Format and attach to Instance > > >>>>>>> * restart instance > > >>>>>>> * newEBSis created > > >>>>>>> * PreviousEBSis orphaned > > > >>>>>>> So each time I have anEBS, but just not the previous one. > > > >>>>>>> Alex Kovalyov wrote: > > >>>>>>>> We've checked your log and found all attach and mount attempts > > >>>>>>>> successful. > > >>>>>>>> Please provide more details. > > > >>>>>>>> On 8 дек, 22:55, Kevin Baker <[email protected]> wrote: > > > >>>>>>>>> So would this apply to anyEBSthat would have been attached on > > >>>>>>>>> startup? > > > >>>>>>>>> I have anEBSon my MySQL instance that has not been > > >>>>>>>>> automatically > > >>>>>>>>> reattaching on restart. > > > >>>>>>>>> Alex Kovalyov wrote: > > > >>>>>>>>>> Confirmed - this was a bug on our side (fixed). > > > >>>>>>>>>> We weren't checking that attachmentSet->status is 'attached' > > >>>>>>>>>> before > > >>>>>>>>>> initiating a volume mount on instance, only checked volume > > >>>>>>>>>> status to > > >>>>>>>>>> be "in-use". > > > >>>>>>>>>> On 8 дек, 11:59, Sam <[email protected]> wrote: > > > >>>>>>>>>>> I have anEBSstorage attached to my app instance. I have > > >>>>>>>>>>> mkdir > > >>>>>>>>>>> fjstorage on application server and then created 2 symlinks to > > >>>>>>>>>>> store > > >>>>>>>>>>> the local files on the server. Then I terminated and > > >>>>>>>>>>> synchronised the > > >>>>>>>>>>> farm and restarted the farm. Everything worked fine. I stopped > > >>>>>>>>>>> the > > >>>>>>>>>>> farm on friday and then restarted it today. > > >>>>>>>>>>> Everything else is fine on the app server but the /fjstorage > > >>>>>>>>>>> directory > > >>>>>>>>>>> whereEBSis to be mounted does not exist. > > >>>>>>>>>>> My settings forEBSin the app server role was to create 10GB > > >>>>>>>>>>> storage > > >>>>>>>>>>> and attach it automatically to /fjstorage directory on the app > > >>>>>>>>>>> server. > > >>>>>>>>>>> I had similar problem before which I was told that they had > > >>>>>>>>>>> fixed it. > > >>>>>>>>>>> This is again happening. > > >>>>>>>>>>> Could you please help urgently as we are going to put an > > >>>>>>>>>>> application > > >>>>>>>>>>> live this week and this sort of problem withEBSstorage > > >>>>>>>>>>> defnitely > > >>>>>>>>>>> won't help. > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "scalr-discuss" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/scalr-discuss?hl=en -~----------~----~----~----~------~----~------~--~---
