Re: Warn about SASI usage and allow to disable them
Can we have a page or a JIRA label which users can see to know why it is experimental. Putting a warning without telling why is not good. But since warning is better than nothing, I am -0 on warn > On Jan 26, 2019, at 9:37 AM, Andrés de la Peña > wrote: > > I agree with Paulo's proposal. I think it will give us a very desirable > homogeneity in how we deal with experimental features. > > I'm +1 to warning, config property, and experimental features (SASI and MV) > disabled by default in trunk. > > These are the explicit votes for now, if I'm counting right: > > - CQL native protocol warning on create SASI index: three +1s, one +0 and > two -0s > - Config property to disable new SASI creation: ten +1s > - New SASI creation disabled by default in trunk: nine +1s and one -0 > - New MV creation disabled by default in trunk: three +1s > > If there are no objections, I'll update the patch by the end of next week. > >> On Mon, 21 Jan 2019 at 19:26, Paulo Motta wrote: >> >> +1 to enable_sasi_indexes flag >> +1 to disabling experimental features by default on 4.0 (SASI and MVs, >> transient replication already disabled) >> >> Regarding the warning on creation of SASI indexes, I think that's a >> user-level warning complimentary to the flag, which is targeted to admins, >> so +1. If people are bothered by this, we could add another flag to disable >> warnings on experimental features, which would be applied to both this and >> MV creation warning (and any other future experimental feature). >> >> I think the warning should be "SASI indexes are experimental and are not >> recommended for production use.", similar to the MV warning added on >> CASSANDRA-13959. >> >> We should open a doc ticket to list limitations of experimental features >> (MVs, SASI, transient replication), but this should probably be out of the >> scope of CASSANDRA-14866. Once we have this doc, we can maybe amend the >> warning to include a link to the doc. >> >> Now that the number of experimental feature flags is growing we should >> perhaps unify all flags in a "experimental features" section on >> cassandra.yaml to allow easily locating them - and a pointer to the >> limitations doc once we have it. >> >> Em qua, 16 de jan de 2019 às 20:18, sankalp kohli >> escreveu: >> >>> If we want to put a warning, we should list in a doc all the open issues >> it >>> has along with explanation of how it can impact. We have a few in the >> first >>> email of this thread but we should put it in a doc for people to know >> what >>> are the issues and if they want to take that risk. >>> >>> >>> >>> On Wed, Jan 16, 2019 at 3:14 PM Brandon Williams >> wrote: >>> Which, if I'm not mistaken, is the goal here? > On Wed, Jan 16, 2019 at 5:12 PM Jeff Jirsa wrote: > > The cost is in how many users you scare away > > -- > Jeff Jirsa > > >> On Jan 16, 2019, at 2:34 PM, Brandon Williams wrote: >> >> Also it costs us nothing to add it. >> >>> On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad < >> j...@jonhaddad.com> > wrote: >>> >>> I'm +1 on the warning for two reasons. >>> A cqlsh warning only applies to those that create the sasi via >>> cqlsh. >>> >>> 1. When people are creating their schemas in development, this is > usually >>> the first step. You use the REPL to figure out what you need, >> then you >>> copy your schema somewhere else. The warning here should prevent >> a lot > of >>> folks from making a serious mistake. >>> >>> 2. It's consistent with how we warn when people try to use materialized >>> views. >>> >>> >>> >>> On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever > wrote: Regarding the warning, we might add it at least in 3.11, since >> for that version the property to enable SASI is going to be present but >> not >>> disabled by default. WDYT? I'm -0 on this. A single line warning in the logs on the sasi creation won't be noticed >>> by many users. A cqlsh warning only applies to those that create the sasi via >>> cqlsh. And we're not talking about patching client drivers to generate a > warning there. So I'd be happy with a yaml comment on the config flag explaining that it's a beta feature and that users should check open tickets and >>> understand current limitations on sasi before using them. regards, Mick >>> - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org >>> >>> -- >>> Jon Haddad >>>
Re: Warn about SASI usage and allow to disable them
I agree with Paulo's proposal. I think it will give us a very desirable homogeneity in how we deal with experimental features. I'm +1 to warning, config property, and experimental features (SASI and MV) disabled by default in trunk. These are the explicit votes for now, if I'm counting right: - CQL native protocol warning on create SASI index: three +1s, one +0 and two -0s - Config property to disable new SASI creation: ten +1s - New SASI creation disabled by default in trunk: nine +1s and one -0 - New MV creation disabled by default in trunk: three +1s If there are no objections, I'll update the patch by the end of next week. On Mon, 21 Jan 2019 at 19:26, Paulo Motta wrote: > +1 to enable_sasi_indexes flag > +1 to disabling experimental features by default on 4.0 (SASI and MVs, > transient replication already disabled) > > Regarding the warning on creation of SASI indexes, I think that's a > user-level warning complimentary to the flag, which is targeted to admins, > so +1. If people are bothered by this, we could add another flag to disable > warnings on experimental features, which would be applied to both this and > MV creation warning (and any other future experimental feature). > > I think the warning should be "SASI indexes are experimental and are not > recommended for production use.", similar to the MV warning added on > CASSANDRA-13959. > > We should open a doc ticket to list limitations of experimental features > (MVs, SASI, transient replication), but this should probably be out of the > scope of CASSANDRA-14866. Once we have this doc, we can maybe amend the > warning to include a link to the doc. > > Now that the number of experimental feature flags is growing we should > perhaps unify all flags in a "experimental features" section on > cassandra.yaml to allow easily locating them - and a pointer to the > limitations doc once we have it. > > Em qua, 16 de jan de 2019 às 20:18, sankalp kohli > escreveu: > > > If we want to put a warning, we should list in a doc all the open issues > it > > has along with explanation of how it can impact. We have a few in the > first > > email of this thread but we should put it in a doc for people to know > what > > are the issues and if they want to take that risk. > > > > > > > > On Wed, Jan 16, 2019 at 3:14 PM Brandon Williams > wrote: > > > > > Which, if I'm not mistaken, is the goal here? > > > > > > On Wed, Jan 16, 2019 at 5:12 PM Jeff Jirsa wrote: > > > > > > > The cost is in how many users you scare away > > > > > > > > -- > > > > Jeff Jirsa > > > > > > > > > > > > > On Jan 16, 2019, at 2:34 PM, Brandon Williams > > > wrote: > > > > > > > > > > Also it costs us nothing to add it. > > > > > > > > > >> On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad < > j...@jonhaddad.com> > > > > wrote: > > > > >> > > > > >> I'm +1 on the warning for two reasons. > > > > >> > > > > >>> A cqlsh warning only applies to those that create the sasi via > > cqlsh. > > > > >> > > > > >> 1. When people are creating their schemas in development, this is > > > > usually > > > > >> the first step. You use the REPL to figure out what you need, > then > > > you > > > > >> copy your schema somewhere else. The warning here should prevent > a > > > lot > > > > of > > > > >> folks from making a serious mistake. > > > > >> > > > > >> 2. It's consistent with how we warn when people try to use > > > materialized > > > > >> views. > > > > >> > > > > >> > > > > >> > > > > >> > > > > >>> On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever > > > > wrote: > > > > >>> > > > > >>> Regarding the warning, we might add it at least in 3.11, since > for > > > that > > > > >>> version the property to enable SASI is going to be present but > not > > > > >> disabled > > > > >>> by default. WDYT? > > > > >>> > > > > >>> > > > > >>> I'm -0 on this. > > > > >>> > > > > >>> A single line warning in the logs on the sasi creation won't be > > > noticed > > > > >> by > > > > >>> many users. > > > > >>> A cqlsh warning only applies to those that create the sasi via > > cqlsh. > > > > >>> And we're not talking about patching client drivers to generate a > > > > warning > > > > >>> there. > > > > >>> > > > > >>> So I'd be happy with a yaml comment on the config flag explaining > > > that > > > > >>> it's a beta feature and that users should check open tickets and > > > > >> understand > > > > >>> current limitations on sasi before using them. > > > > >>> > > > > >>> regards, > > > > >>> Mick > > > > >>> > > > > >>> > > - > > > > >>> To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > > > >>> For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > >>> > > > > >>> > > > > >> > > > > >> -- > > > > >> Jon Haddad > > > > >> http://www.rustyrazorblade.com > > > > >> twitter: rustyrazorblade > > > > >> > > > > > > > > - > > > > To unsubscribe,
Re: Warn about SASI usage and allow to disable them
+1 to enable_sasi_indexes flag +1 to disabling experimental features by default on 4.0 (SASI and MVs, transient replication already disabled) Regarding the warning on creation of SASI indexes, I think that's a user-level warning complimentary to the flag, which is targeted to admins, so +1. If people are bothered by this, we could add another flag to disable warnings on experimental features, which would be applied to both this and MV creation warning (and any other future experimental feature). I think the warning should be "SASI indexes are experimental and are not recommended for production use.", similar to the MV warning added on CASSANDRA-13959. We should open a doc ticket to list limitations of experimental features (MVs, SASI, transient replication), but this should probably be out of the scope of CASSANDRA-14866. Once we have this doc, we can maybe amend the warning to include a link to the doc. Now that the number of experimental feature flags is growing we should perhaps unify all flags in a "experimental features" section on cassandra.yaml to allow easily locating them - and a pointer to the limitations doc once we have it. Em qua, 16 de jan de 2019 às 20:18, sankalp kohli escreveu: > If we want to put a warning, we should list in a doc all the open issues it > has along with explanation of how it can impact. We have a few in the first > email of this thread but we should put it in a doc for people to know what > are the issues and if they want to take that risk. > > > > On Wed, Jan 16, 2019 at 3:14 PM Brandon Williams wrote: > > > Which, if I'm not mistaken, is the goal here? > > > > On Wed, Jan 16, 2019 at 5:12 PM Jeff Jirsa wrote: > > > > > The cost is in how many users you scare away > > > > > > -- > > > Jeff Jirsa > > > > > > > > > > On Jan 16, 2019, at 2:34 PM, Brandon Williams > > wrote: > > > > > > > > Also it costs us nothing to add it. > > > > > > > >> On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad > > > wrote: > > > >> > > > >> I'm +1 on the warning for two reasons. > > > >> > > > >>> A cqlsh warning only applies to those that create the sasi via > cqlsh. > > > >> > > > >> 1. When people are creating their schemas in development, this is > > > usually > > > >> the first step. You use the REPL to figure out what you need, then > > you > > > >> copy your schema somewhere else. The warning here should prevent a > > lot > > > of > > > >> folks from making a serious mistake. > > > >> > > > >> 2. It's consistent with how we warn when people try to use > > materialized > > > >> views. > > > >> > > > >> > > > >> > > > >> > > > >>> On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever > > > wrote: > > > >>> > > > >>> Regarding the warning, we might add it at least in 3.11, since for > > that > > > >>> version the property to enable SASI is going to be present but not > > > >> disabled > > > >>> by default. WDYT? > > > >>> > > > >>> > > > >>> I'm -0 on this. > > > >>> > > > >>> A single line warning in the logs on the sasi creation won't be > > noticed > > > >> by > > > >>> many users. > > > >>> A cqlsh warning only applies to those that create the sasi via > cqlsh. > > > >>> And we're not talking about patching client drivers to generate a > > > warning > > > >>> there. > > > >>> > > > >>> So I'd be happy with a yaml comment on the config flag explaining > > that > > > >>> it's a beta feature and that users should check open tickets and > > > >> understand > > > >>> current limitations on sasi before using them. > > > >>> > > > >>> regards, > > > >>> Mick > > > >>> > > > >>> > - > > > >>> To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > > >>> For additional commands, e-mail: dev-h...@cassandra.apache.org > > > >>> > > > >>> > > > >> > > > >> -- > > > >> Jon Haddad > > > >> http://www.rustyrazorblade.com > > > >> twitter: rustyrazorblade > > > >> > > > > > > - > > > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > > For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > > > > > >
Re: Warn about SASI usage and allow to disable them
Which, if I'm not mistaken, is the goal here? On Wed, Jan 16, 2019 at 5:12 PM Jeff Jirsa wrote: > The cost is in how many users you scare away > > -- > Jeff Jirsa > > > > On Jan 16, 2019, at 2:34 PM, Brandon Williams wrote: > > > > Also it costs us nothing to add it. > > > >> On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad > wrote: > >> > >> I'm +1 on the warning for two reasons. > >> > >>> A cqlsh warning only applies to those that create the sasi via cqlsh. > >> > >> 1. When people are creating their schemas in development, this is > usually > >> the first step. You use the REPL to figure out what you need, then you > >> copy your schema somewhere else. The warning here should prevent a lot > of > >> folks from making a serious mistake. > >> > >> 2. It's consistent with how we warn when people try to use materialized > >> views. > >> > >> > >> > >> > >>> On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever > wrote: > >>> > >>> Regarding the warning, we might add it at least in 3.11, since for that > >>> version the property to enable SASI is going to be present but not > >> disabled > >>> by default. WDYT? > >>> > >>> > >>> I'm -0 on this. > >>> > >>> A single line warning in the logs on the sasi creation won't be noticed > >> by > >>> many users. > >>> A cqlsh warning only applies to those that create the sasi via cqlsh. > >>> And we're not talking about patching client drivers to generate a > warning > >>> there. > >>> > >>> So I'd be happy with a yaml comment on the config flag explaining that > >>> it's a beta feature and that users should check open tickets and > >> understand > >>> current limitations on sasi before using them. > >>> > >>> regards, > >>> Mick > >>> > >>> - > >>> To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > >>> For additional commands, e-mail: dev-h...@cassandra.apache.org > >>> > >>> > >> > >> -- > >> Jon Haddad > >> http://www.rustyrazorblade.com > >> twitter: rustyrazorblade > >> > > - > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > For additional commands, e-mail: dev-h...@cassandra.apache.org > >
Re: Warn about SASI usage and allow to disable them
The cost is in how many users you scare away -- Jeff Jirsa > On Jan 16, 2019, at 2:34 PM, Brandon Williams wrote: > > Also it costs us nothing to add it. > >> On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad wrote: >> >> I'm +1 on the warning for two reasons. >> >>> A cqlsh warning only applies to those that create the sasi via cqlsh. >> >> 1. When people are creating their schemas in development, this is usually >> the first step. You use the REPL to figure out what you need, then you >> copy your schema somewhere else. The warning here should prevent a lot of >> folks from making a serious mistake. >> >> 2. It's consistent with how we warn when people try to use materialized >> views. >> >> >> >> >>> On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever wrote: >>> >>> Regarding the warning, we might add it at least in 3.11, since for that >>> version the property to enable SASI is going to be present but not >> disabled >>> by default. WDYT? >>> >>> >>> I'm -0 on this. >>> >>> A single line warning in the logs on the sasi creation won't be noticed >> by >>> many users. >>> A cqlsh warning only applies to those that create the sasi via cqlsh. >>> And we're not talking about patching client drivers to generate a warning >>> there. >>> >>> So I'd be happy with a yaml comment on the config flag explaining that >>> it's a beta feature and that users should check open tickets and >> understand >>> current limitations on sasi before using them. >>> >>> regards, >>> Mick >>> >>> - >>> To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org >>> For additional commands, e-mail: dev-h...@cassandra.apache.org >>> >>> >> >> -- >> Jon Haddad >> http://www.rustyrazorblade.com >> twitter: rustyrazorblade >> - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org
Re: Warn about SASI usage and allow to disable them
Also it costs us nothing to add it. On Wed, Jan 16, 2019 at 4:29 PM Jonathan Haddad wrote: > I'm +1 on the warning for two reasons. > > > A cqlsh warning only applies to those that create the sasi via cqlsh. > > 1. When people are creating their schemas in development, this is usually > the first step. You use the REPL to figure out what you need, then you > copy your schema somewhere else. The warning here should prevent a lot of > folks from making a serious mistake. > > 2. It's consistent with how we warn when people try to use materialized > views. > > > > > On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever wrote: > > > Regarding the warning, we might add it at least in 3.11, since for that > > version the property to enable SASI is going to be present but not > disabled > > by default. WDYT? > > > > > > I'm -0 on this. > > > > A single line warning in the logs on the sasi creation won't be noticed > by > > many users. > > A cqlsh warning only applies to those that create the sasi via cqlsh. > > And we're not talking about patching client drivers to generate a warning > > there. > > > > So I'd be happy with a yaml comment on the config flag explaining that > > it's a beta feature and that users should check open tickets and > understand > > current limitations on sasi before using them. > > > > regards, > > Mick > > > > - > > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > > > -- > Jon Haddad > http://www.rustyrazorblade.com > twitter: rustyrazorblade >
Re: Warn about SASI usage and allow to disable them
I'm +1 on the warning for two reasons. > A cqlsh warning only applies to those that create the sasi via cqlsh. 1. When people are creating their schemas in development, this is usually the first step. You use the REPL to figure out what you need, then you copy your schema somewhere else. The warning here should prevent a lot of folks from making a serious mistake. 2. It's consistent with how we warn when people try to use materialized views. On Wed, Jan 16, 2019 at 2:15 PM Mick Semb Wever wrote: > Regarding the warning, we might add it at least in 3.11, since for that > version the property to enable SASI is going to be present but not disabled > by default. WDYT? > > > I'm -0 on this. > > A single line warning in the logs on the sasi creation won't be noticed by > many users. > A cqlsh warning only applies to those that create the sasi via cqlsh. > And we're not talking about patching client drivers to generate a warning > there. > > So I'd be happy with a yaml comment on the config flag explaining that > it's a beta feature and that users should check open tickets and understand > current limitations on sasi before using them. > > regards, > Mick > > - > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > For additional commands, e-mail: dev-h...@cassandra.apache.org > > -- Jon Haddad http://www.rustyrazorblade.com twitter: rustyrazorblade
Re: Warn about SASI usage and allow to disable them
Regarding the warning, we might add it at least in 3.11, since for that version the property to enable SASI is going to be present but not disabled by default. WDYT? I'm -0 on this. A single line warning in the logs on the sasi creation won't be noticed by many users. A cqlsh warning only applies to those that create the sasi via cqlsh. And we're not talking about patching client drivers to generate a warning there. So I'd be happy with a yaml comment on the config flag explaining that it's a beta feature and that users should check open tickets and understand current limitations on sasi before using them. regards, Mick - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org
Re: Warn about SASI usage and allow to disable them
Thanks for the feedback. It seems we agree to the config property, disabled by default in trunk. Regarding the warning, we might add it at least in 3.11, since for that version the property to enable SASI is going to be present but not disabled by default. WDYT? On Tue, 15 Jan 2019 at 11:24, Benjamin Lerer wrote: > +1 on config. +1 on disabling by default > > On Tue, Jan 15, 2019 at 6:51 AM Jeff Jirsa wrote: > > > +1 on config > > -0 on warning > > -0 on disabling by default > > > > > > -- > > Jeff Jirsa > > > > > > > On Jan 14, 2019, at 9:22 PM, Taylor Cressy > > wrote: > > > > > > +1 on config. +1 on disabling. > > > > > > +1 on applying it to materialized views as well. > > > > > >> On Jan 14, 2019, at 17:29, Joshua McKenzie > > wrote: > > >> > > >> +1 on config change, +1 on disabling, and so long as the comments make > > the > > >> limitations and risks extremely clear, I'm fine w/out the client > > warning. > > >> > > >> On Mon, Jan 14, 2019 at 12:28 PM Andrés de la Peña < > > a.penya.gar...@gmail.com> > > >> wrote: > > >> > > >>> I mean disabling the creation of new SASI indices with CREATE INDEX > > >>> statement, the existing indexes would continue working. The CQL > client > > >>> warning will be thrown with that creation statement as well (if they > > are > > >>> enabled). > > >>> > > On Mon, 14 Jan 2019 at 20:18, Jeff Jirsa wrote: > > > > When we say disable, do you mean disable creation of new SASI > > indices, or > > disable using existing ones? I assume it's just creation of new? > > > > On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < > > a.penya.gar...@gmail.com> > > wrote: > > > > > Hello all, > > > > > > It is my understanding that SASI is still to be considered an > > > experimental/beta feature, and they apparently are not being very > > actively > > > developed. Some higlighted problems in SASI are: > > > > > > - OOMs during flush, as it is described in CASSANDRA-12662 > > > - General secondary index consistency problems described in > > CASSANDRA-8272. > > > There is a pending-review patch addressing the problem for regular > > 2i. > > > However, the proposed solution is based on indexing tombstones. > SASI > > > doesn't index tombstones, so it wouldn't be enterely trivial to > > extend > > the > > > approach to SASI. > > > - Probably insufficient testing. As far as I know, we don't have a > > >>> single > > > dtest for SASI nor tests dealing with large SSTables. > > > > > > Similarly to what CASSANDRA-13959 did with materialized views, > > > CASSANDRA-14866 aims to throw a native protocol warning about SASI > > > experimental state, and to add a config property to disable them. > > >>> Perhaps > > > this property could be disabled by default in trunk. This should > > raise > > > awareness about SASI maturity until we let them in a more stable > > state. > > > > > > The purpose for this thread is discussing whether we want to add > this > > > warning, the config property and, more controversially, if we want > to > > >>> set > > > SASI as disabled by default in trunk. > > > > > > WDYT? > > > > > > > >>> > > > > > > - > > > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > > For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > > > > - > > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > >
Re: Warn about SASI usage and allow to disable them
+1 on config -0 on warning -0 on disabling by default -- Jeff Jirsa > On Jan 14, 2019, at 9:22 PM, Taylor Cressy wrote: > > +1 on config. +1 on disabling. > > +1 on applying it to materialized views as well. > >> On Jan 14, 2019, at 17:29, Joshua McKenzie wrote: >> >> +1 on config change, +1 on disabling, and so long as the comments make the >> limitations and risks extremely clear, I'm fine w/out the client warning. >> >> On Mon, Jan 14, 2019 at 12:28 PM Andrés de la Peña >> wrote: >> >>> I mean disabling the creation of new SASI indices with CREATE INDEX >>> statement, the existing indexes would continue working. The CQL client >>> warning will be thrown with that creation statement as well (if they are >>> enabled). >>> On Mon, 14 Jan 2019 at 20:18, Jeff Jirsa wrote: When we say disable, do you mean disable creation of new SASI indices, or disable using existing ones? I assume it's just creation of new? On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < a.penya.gar...@gmail.com> wrote: > Hello all, > > It is my understanding that SASI is still to be considered an > experimental/beta feature, and they apparently are not being very actively > developed. Some higlighted problems in SASI are: > > - OOMs during flush, as it is described in CASSANDRA-12662 > - General secondary index consistency problems described in CASSANDRA-8272. > There is a pending-review patch addressing the problem for regular 2i. > However, the proposed solution is based on indexing tombstones. SASI > doesn't index tombstones, so it wouldn't be enterely trivial to extend the > approach to SASI. > - Probably insufficient testing. As far as I know, we don't have a >>> single > dtest for SASI nor tests dealing with large SSTables. > > Similarly to what CASSANDRA-13959 did with materialized views, > CASSANDRA-14866 aims to throw a native protocol warning about SASI > experimental state, and to add a config property to disable them. >>> Perhaps > this property could be disabled by default in trunk. This should raise > awareness about SASI maturity until we let them in a more stable state. > > The purpose for this thread is discussing whether we want to add this > warning, the config property and, more controversially, if we want to >>> set > SASI as disabled by default in trunk. > > WDYT? > >>> > > - > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > For additional commands, e-mail: dev-h...@cassandra.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org
Re: Warn about SASI usage and allow to disable them
+1 on config. +1 on disabling. +1 on applying it to materialized views as well. > On Jan 14, 2019, at 17:29, Joshua McKenzie wrote: > > +1 on config change, +1 on disabling, and so long as the comments make the > limitations and risks extremely clear, I'm fine w/out the client warning. > > On Mon, Jan 14, 2019 at 12:28 PM Andrés de la Peña > wrote: > >> I mean disabling the creation of new SASI indices with CREATE INDEX >> statement, the existing indexes would continue working. The CQL client >> warning will be thrown with that creation statement as well (if they are >> enabled). >> >>> On Mon, 14 Jan 2019 at 20:18, Jeff Jirsa wrote: >>> >>> When we say disable, do you mean disable creation of new SASI indices, or >>> disable using existing ones? I assume it's just creation of new? >>> >>> On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < >>> a.penya.gar...@gmail.com> >>> wrote: >>> Hello all, It is my understanding that SASI is still to be considered an experimental/beta feature, and they apparently are not being very >>> actively developed. Some higlighted problems in SASI are: - OOMs during flush, as it is described in CASSANDRA-12662 - General secondary index consistency problems described in >>> CASSANDRA-8272. There is a pending-review patch addressing the problem for regular 2i. However, the proposed solution is based on indexing tombstones. SASI doesn't index tombstones, so it wouldn't be enterely trivial to extend >>> the approach to SASI. - Probably insufficient testing. As far as I know, we don't have a >> single dtest for SASI nor tests dealing with large SSTables. Similarly to what CASSANDRA-13959 did with materialized views, CASSANDRA-14866 aims to throw a native protocol warning about SASI experimental state, and to add a config property to disable them. >> Perhaps this property could be disabled by default in trunk. This should raise awareness about SASI maturity until we let them in a more stable state. The purpose for this thread is discussing whether we want to add this warning, the config property and, more controversially, if we want to >> set SASI as disabled by default in trunk. WDYT? >>> >> - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org
Re: Warn about SASI usage and allow to disable them
+1 on config change, +1 on disabling, and so long as the comments make the limitations and risks extremely clear, I'm fine w/out the client warning. On Mon, Jan 14, 2019 at 12:28 PM Andrés de la Peña wrote: > I mean disabling the creation of new SASI indices with CREATE INDEX > statement, the existing indexes would continue working. The CQL client > warning will be thrown with that creation statement as well (if they are > enabled). > > On Mon, 14 Jan 2019 at 20:18, Jeff Jirsa wrote: > > > When we say disable, do you mean disable creation of new SASI indices, or > > disable using existing ones? I assume it's just creation of new? > > > > On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < > > a.penya.gar...@gmail.com> > > wrote: > > > > > Hello all, > > > > > > It is my understanding that SASI is still to be considered an > > > experimental/beta feature, and they apparently are not being very > > actively > > > developed. Some higlighted problems in SASI are: > > > > > > - OOMs during flush, as it is described in CASSANDRA-12662 > > > - General secondary index consistency problems described in > > CASSANDRA-8272. > > > There is a pending-review patch addressing the problem for regular 2i. > > > However, the proposed solution is based on indexing tombstones. SASI > > > doesn't index tombstones, so it wouldn't be enterely trivial to extend > > the > > > approach to SASI. > > > - Probably insufficient testing. As far as I know, we don't have a > single > > > dtest for SASI nor tests dealing with large SSTables. > > > > > > Similarly to what CASSANDRA-13959 did with materialized views, > > > CASSANDRA-14866 aims to throw a native protocol warning about SASI > > > experimental state, and to add a config property to disable them. > Perhaps > > > this property could be disabled by default in trunk. This should raise > > > awareness about SASI maturity until we let them in a more stable state. > > > > > > The purpose for this thread is discussing whether we want to add this > > > warning, the config property and, more controversially, if we want to > set > > > SASI as disabled by default in trunk. > > > > > > WDYT? > > > > > >
Re: Warn about SASI usage and allow to disable them
I mean disabling the creation of new SASI indices with CREATE INDEX statement, the existing indexes would continue working. The CQL client warning will be thrown with that creation statement as well (if they are enabled). On Mon, 14 Jan 2019 at 20:18, Jeff Jirsa wrote: > When we say disable, do you mean disable creation of new SASI indices, or > disable using existing ones? I assume it's just creation of new? > > On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < > a.penya.gar...@gmail.com> > wrote: > > > Hello all, > > > > It is my understanding that SASI is still to be considered an > > experimental/beta feature, and they apparently are not being very > actively > > developed. Some higlighted problems in SASI are: > > > > - OOMs during flush, as it is described in CASSANDRA-12662 > > - General secondary index consistency problems described in > CASSANDRA-8272. > > There is a pending-review patch addressing the problem for regular 2i. > > However, the proposed solution is based on indexing tombstones. SASI > > doesn't index tombstones, so it wouldn't be enterely trivial to extend > the > > approach to SASI. > > - Probably insufficient testing. As far as I know, we don't have a single > > dtest for SASI nor tests dealing with large SSTables. > > > > Similarly to what CASSANDRA-13959 did with materialized views, > > CASSANDRA-14866 aims to throw a native protocol warning about SASI > > experimental state, and to add a config property to disable them. Perhaps > > this property could be disabled by default in trunk. This should raise > > awareness about SASI maturity until we let them in a more stable state. > > > > The purpose for this thread is discussing whether we want to add this > > warning, the config property and, more controversially, if we want to set > > SASI as disabled by default in trunk. > > > > WDYT? > > >
Re: Warn about SASI usage and allow to disable them
When we say disable, do you mean disable creation of new SASI indices, or disable using existing ones? I assume it's just creation of new? On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña wrote: > Hello all, > > It is my understanding that SASI is still to be considered an > experimental/beta feature, and they apparently are not being very actively > developed. Some higlighted problems in SASI are: > > - OOMs during flush, as it is described in CASSANDRA-12662 > - General secondary index consistency problems described in CASSANDRA-8272. > There is a pending-review patch addressing the problem for regular 2i. > However, the proposed solution is based on indexing tombstones. SASI > doesn't index tombstones, so it wouldn't be enterely trivial to extend the > approach to SASI. > - Probably insufficient testing. As far as I know, we don't have a single > dtest for SASI nor tests dealing with large SSTables. > > Similarly to what CASSANDRA-13959 did with materialized views, > CASSANDRA-14866 aims to throw a native protocol warning about SASI > experimental state, and to add a config property to disable them. Perhaps > this property could be disabled by default in trunk. This should raise > awareness about SASI maturity until we let them in a more stable state. > > The purpose for this thread is discussing whether we want to add this > warning, the config property and, more controversially, if we want to set > SASI as disabled by default in trunk. > > WDYT? >
Re: Warn about SASI usage and allow to disable them
I amend to +1 everything except warning, which I'm +0 on. On Mon, Jan 14, 2019 at 1:59 PM Caleb Rackliffe wrote: > +1 to config and disable > > On Mon, Jan 14, 2019, 1:54 PM Mick Semb Wever > > > > > > > The purpose for this thread is discussing whether we want to add this > > > warning, the config property and, more controversially, if we want to > set > > > SASI as disabled by default in trunk. > > > > > > I'm +1 on everything, except the warning. > > > > I think if we add the config property and it's disabled in trunk then > > we're done enough. > > Existing users should be ok. > > > > regards, > > Mick > > > > - > > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > > For additional commands, e-mail: dev-h...@cassandra.apache.org > > > > >
Re: Warn about SASI usage and allow to disable them
+1 to config and disable On Mon, Jan 14, 2019, 1:54 PM Mick Semb Wever > > > The purpose for this thread is discussing whether we want to add this > > warning, the config property and, more controversially, if we want to set > > SASI as disabled by default in trunk. > > > I'm +1 on everything, except the warning. > > I think if we add the config property and it's disabled in trunk then > we're done enough. > Existing users should be ok. > > regards, > Mick > > - > To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org > For additional commands, e-mail: dev-h...@cassandra.apache.org > >
Re: Warn about SASI usage and allow to disable them
> The purpose for this thread is discussing whether we want to add this > warning, the config property and, more controversially, if we want to set > SASI as disabled by default in trunk. I'm +1 on everything, except the warning. I think if we add the config property and it's disabled in trunk then we're done enough. Existing users should be ok. regards, Mick - To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org For additional commands, e-mail: dev-h...@cassandra.apache.org
Re: Warn about SASI usage and allow to disable them
+1 to warn, config, and disable. On Mon, Jan 14, 2019 at 1:45 PM Jonathan Haddad wrote: > I'm very much in favor of a warning, and I lean towards disabling them (and > MVs, while we're at it) by default as well. > > I've seen both features be the death of clusters, and are a major risk for > teams that are brand new to Cassandra. > > > > On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña < > a.penya.gar...@gmail.com> > wrote: > > > Hello all, > > > > It is my understanding that SASI is still to be considered an > > experimental/beta feature, and they apparently are not being very > actively > > developed. Some higlighted problems in SASI are: > > > > - OOMs during flush, as it is described in CASSANDRA-12662 > > - General secondary index consistency problems described in > CASSANDRA-8272. > > There is a pending-review patch addressing the problem for regular 2i. > > However, the proposed solution is based on indexing tombstones. SASI > > doesn't index tombstones, so it wouldn't be enterely trivial to extend > the > > approach to SASI. > > - Probably insufficient testing. As far as I know, we don't have a single > > dtest for SASI nor tests dealing with large SSTables. > > > > Similarly to what CASSANDRA-13959 did with materialized views, > > CASSANDRA-14866 aims to throw a native protocol warning about SASI > > experimental state, and to add a config property to disable them. Perhaps > > this property could be disabled by default in trunk. This should raise > > awareness about SASI maturity until we let them in a more stable state. > > > > The purpose for this thread is discussing whether we want to add this > > warning, the config property and, more controversially, if we want to set > > SASI as disabled by default in trunk. > > > > WDYT? > > > > > -- > Jon Haddad > http://www.rustyrazorblade.com > twitter: rustyrazorblade >
Re: Warn about SASI usage and allow to disable them
I'm very much in favor of a warning, and I lean towards disabling them (and MVs, while we're at it) by default as well. I've seen both features be the death of clusters, and are a major risk for teams that are brand new to Cassandra. On Mon, Jan 14, 2019 at 11:19 AM Andrés de la Peña wrote: > Hello all, > > It is my understanding that SASI is still to be considered an > experimental/beta feature, and they apparently are not being very actively > developed. Some higlighted problems in SASI are: > > - OOMs during flush, as it is described in CASSANDRA-12662 > - General secondary index consistency problems described in CASSANDRA-8272. > There is a pending-review patch addressing the problem for regular 2i. > However, the proposed solution is based on indexing tombstones. SASI > doesn't index tombstones, so it wouldn't be enterely trivial to extend the > approach to SASI. > - Probably insufficient testing. As far as I know, we don't have a single > dtest for SASI nor tests dealing with large SSTables. > > Similarly to what CASSANDRA-13959 did with materialized views, > CASSANDRA-14866 aims to throw a native protocol warning about SASI > experimental state, and to add a config property to disable them. Perhaps > this property could be disabled by default in trunk. This should raise > awareness about SASI maturity until we let them in a more stable state. > > The purpose for this thread is discussing whether we want to add this > warning, the config property and, more controversially, if we want to set > SASI as disabled by default in trunk. > > WDYT? > -- Jon Haddad http://www.rustyrazorblade.com twitter: rustyrazorblade
Warn about SASI usage and allow to disable them
Hello all, It is my understanding that SASI is still to be considered an experimental/beta feature, and they apparently are not being very actively developed. Some higlighted problems in SASI are: - OOMs during flush, as it is described in CASSANDRA-12662 - General secondary index consistency problems described in CASSANDRA-8272. There is a pending-review patch addressing the problem for regular 2i. However, the proposed solution is based on indexing tombstones. SASI doesn't index tombstones, so it wouldn't be enterely trivial to extend the approach to SASI. - Probably insufficient testing. As far as I know, we don't have a single dtest for SASI nor tests dealing with large SSTables. Similarly to what CASSANDRA-13959 did with materialized views, CASSANDRA-14866 aims to throw a native protocol warning about SASI experimental state, and to add a config property to disable them. Perhaps this property could be disabled by default in trunk. This should raise awareness about SASI maturity until we let them in a more stable state. The purpose for this thread is discussing whether we want to add this warning, the config property and, more controversially, if we want to set SASI as disabled by default in trunk. WDYT?
