Add the Hypertable.Failover.Quorum.Percentage=100 property as mentioned in the email I just sent out. That will prevent the issue from occurring again on restart.
- Doug On Fri, Feb 22, 2013 at 10:44 AM, Alex Kashirin <[email protected]>wrote: > it is on the 0.9.7.0.c7c043a Version by the way. > hope, nothing will happen on my Cluster till the upcoming release. > Alex > > On Friday, February 22, 2013 8:37:54 PM UTC+2, Doug Judd wrote: > >> I'm glad to hear that failover worked for you! :) Thanks for offering >> to help, but I've pinned down the issue. The problem has to do with the >> Master proceeding with a recovery operation on startup when it should not. >> For existing databases that take some time to come up, if there are any >> stragglers, the Master assumes that they have failed and begins recovery >> procedure. Anyway, the fix is fairly straightforward. We'll do a 0.9.7.1 >> release early next week. >> >> - Doug >> >> >> On Fri, Feb 22, 2013 at 10:26 AM, Alex Kashirin <[email protected]>wrote: >> >>> have just deleted the location, started the RangeServers, >>> They are running , and range count are been assigned. >>> >>> Follow: ( >>> https://groups.google.com/forum/?fromgroups=#!topic/hypertable-user/SZU7dvI4mGg >>> ) >>> >>> just letting you know. (while, the ranges are on a new location) >>> >>> www.Epidemic-Online.com >>> Kashirin Alex >>> >>> On Friday, February 22, 2013 7:27:13 PM UTC+2, Alex Kashirin wrote: >>>> >>>> a Marked as removed status! - at the RangeServer log >>>> >>>> (the HT is running for me, so, to touch anything about having all the >>>> servers running at tis moment is not urgent) >>>> >>>> On Friday, February 22, 2013 7:22:08 PM UTC+2, Alex Kashirin wrote: >>>>> >>>>> don't know if it is the same problem, >>>>> >>>>> server was taken off-line for some-time (running range and hyperspace) >>>>> >>>>> by Monitoring, the RangeServer stayed at no response, >>>>> with the server back on to networking the rangeserver status did not >>>>> change. >>>>> >>>>> restarted the cluster, just TWO RangeServer are not showing up, >>>>> (while, it is running PID) - the HT is working and the Ranges Count have >>>>> re-balanced >>>>> >>>>> >>>>> DOUG, if you want to evaluate the problem , email me I'll send you the >>>>> details to the servers, OR do you need any of the logs. >>>>> >>>>> thanks, >>>>> Alex >>>>> >>>>> >>>>> >>>>> On Friday, February 22, 2013 6:50:52 PM UTC+2, Doug Judd wrote: >>>>>> >>>>>> I've reproduced the problem that a few people have reported on the >>>>>> mailing list when upgrading to 0.9.7.0. We'll fix it ASAP and release >>>>>> 0.9.7.1 early next week. I apologize for any problems this may have >>>>>> caused >>>>>> you. >>>>>> >>>>>> - Doug >>>>>> >>>>>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Hypertable User" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To post to this group, send email to [email protected]. >>> >>> Visit this group at http://groups.google.com/group/hypertable-user?hl=en >>> . >>> For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >>> >> >> >> >> -- >> Doug Judd >> CEO, Hypertable Inc. >> > -- > You received this message because you are subscribed to the Google Groups > "Hypertable User" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/hypertable-user?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- Doug Judd CEO, Hypertable Inc. -- You received this message because you are subscribed to the Google Groups "Hypertable Development" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/hypertable-dev?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
