Hi, This is a single database and it's only 1/100th of it's intended/required size.
I wish that only the dirty/delta would be indexed in a scenario like this, instead of the whole database. The .wal file grow rather large (7-10 files totaling several gigs) under load and the shutdown takes for ever. Sometimes we can not afford such a long "bounce" time. Regards, -Stefan On Tuesday, 7 October 2014 09:24:04 UTC, Andrey Lomakin wrote: > > Hi Fabrizio , Stefan . > > Stefan do you have several databases or single one ? > On Oct 7, 2014 11:56 AM, "Fabrizio Fortino" <[email protected] > <javascript:>> wrote: > >> Hi Andrey, >> >> I am experiencing a similar problem using OrientDB 1.7.4 on Linux CentOS. >> I have around 100 databases and the shutdown process takes more than 10 >> minutes (6/7 seconds for each db). >> >> I moved the server with all the databases on my local machine (Mac OS X) >> and the shutdown here takes seconds. >> >> @Stefan what OS do you use? >> >> Can this problem be related to the environment? >> >> Hope it helps, >> Fabrizio >> >> On Tuesday, October 7, 2014 6:33:52 AM UTC+1, Andrey Lomakin wrote: >>> >>> HI Stefan :-), >>> >>> Index rebuild is needed if you work in non tx mode in such case you need >>> to rebuild it after crash because indexes and documents can be out of sync. >>> But if you work in tx mode indexes rebuild can be switched off. >>> >>> We will fix even this index rebuild in 2.1 but still if you work in tx >>> mode only you can use following settings: >>> index.auto.rebuildAfterNotSoftClose = false >>> index.txMode = FULL >>> >>> But as I can see it you do hard close because shotdown takes a lot of >>> time, so may be would be better to speed up shutdown instead. I am asking >>> because those settings decrease insertion speed. >>> >>> >>> On Mon, Oct 6, 2014 at 1:48 PM, <[email protected]> wrote: >>> >>>> bump. >>>> >>>> This should matter to anyone using ODB in product. >>>> >>>> Regards, >>>> -Stefán >>>> >>>> >>>> On Wednesday, 1 October 2014 20:20:08 UTC, [email protected] >>>> wrote: >>>>> >>>>> Hi, >>>>> >>>>> I'm using an embedded server (2.0-M2 plocal) an it has happened that >>>>> the application has been force/hard-closed along with ODB. >>>>> >>>>> I know that I can start it again without rebuilding indexes after >>>>> hard-close but the though of unindexed records is not appealing. >>>>> >>>>> I currently have a small graph (2.7g) with a single UniqueHash index >>>>> that is being rebuilt from scratch and it's taking forever (40min now and >>>>> counting). >>>>> >>>>> Can someone please walk me through the rationale behind building the >>>>> complete index again under these circumstances and what it would take to >>>>> index only the possible delta (from wal, I presume). >>>>> >>>>> Regards, >>>>> -Stefán >>>>> >>>> -- >>>> >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "OrientDB" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> >>> >>> >>> -- >>> Best regards, >>> Andrey Lomakin. >>> >>> Orient Technologies >>> the Company behind OrientDB >>> >>> -- >> >> --- >> You received this message because you are subscribed to the Google Groups >> "OrientDB" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected] <javascript:>. >> For more options, visit https://groups.google.com/d/optout. >> > -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
