We're wrapping up development for 1.6.0 and could use all the help we can get testing this release. Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x version for Tuesday or Wednesday next week, and release 1.6.0 either Friday 7/30 or Monday 8/2.
Major features to test: - Replica Sets : ready for testing. some loose ends we're still fixing, but testers would be greatly appreciate - Sharding : lots of stability/maturity improvements - testers would be great
I see that replica pairs are going away in 1.6, but is master/slave
still supposed to work? I can't get a slave to replicate using 1.5.6
as master. It starts up and begins to replicate, but seems to get
stuck on the first DB it hits (which is very small). If I switch to
using 1.4.4 as master, a 1.5.6 slave replicates fine.
The logs show it creating a bunch of temporary collections from the
map-reduces we run, and I just see a bunch of this:
[replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was
empty, skipping ahead local.sources
[replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425
operations
[replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38
2010 4c4f4da6:6
Any suggestions?
Thanks,
Ian
On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
> We're wrapping up development for 1.6.0 and could use all the help we
> can get testing this release.
> Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x
> version for Tuesday or Wednesday next week, and release 1.6.0 either
> Friday 7/30 or Monday 8/2.
> Major features to test:
> - Replica Sets : ready for testing. some loose ends we're still
> fixing, but testers would be greatly appreciate
> - Sharding : lots of stability/maturity improvements - testers would be great
Seems like master/slave still works, but the behavior is different. I
reverted the master to 1.4.4 and started up the slave on 1.5.6. It
did a full resync of around 100GB in a few hours. I then upgraded the
master back to 1.5.6, and the slave seems to still be replicating, but
instead of staying within a second or two of the master, it waits a
full minute before grabbing the most recent changes, so it could be up
to a minute behind at any point. Is this expected behavior?
- Ian
On Jul 27, 4:21 pm, iragsdale <ian.ragsd...@gmail.com> wrote:
> I see that replica pairs are going away in 1.6, but is master/slave
> still supposed to work? I can't get a slave to replicate using 1.5.6
> as master. It starts up and begins to replicate, but seems to get
> stuck on the first DB it hits (which is very small). If I switch to
> using 1.4.4 as master, a 1.5.6 slave replicates fine.
> The logs show it creating a bunch of temporary collections from the
> map-reduces we run, and I just see a bunch of this:
> [replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was
> empty, skipping ahead local.sources
> [replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425
> operations
> [replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38
> 2010 4c4f4da6:6
> Any suggestions?
> Thanks,
> Ian
> On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
> > MongoDB 1.5.6 is out and ready for testing.
> > We're wrapping up development for 1.6.0 and could use all the help we
> > can get testing this release.
> > Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x
> > version for Tuesday or Wednesday next week, and release 1.6.0 either
> > Friday 7/30 or Monday 8/2.
> > Major features to test:
> > - Replica Sets : ready for testing. some loose ends we're still
> > fixing, but testers would be greatly appreciate
> > - Sharding : lots of stability/maturity improvements - testers would be great
On Tue, Jul 27, 2010 at 11:28 PM, iragsdale <ian.ragsd...@gmail.com> wrote: > Seems like master/slave still works, but the behavior is different. I > reverted the master to 1.4.4 and started up the slave on 1.5.6. It > did a full resync of around 100GB in a few hours. I then upgraded the > master back to 1.5.6, and the slave seems to still be replicating, but > instead of staying within a second or two of the master, it waits a > full minute before grabbing the most recent changes, so it could be up > to a minute behind at any point. Is this expected behavior?
> - Ian
> On Jul 27, 4:21 pm, iragsdale <ian.ragsd...@gmail.com> wrote: >> I see that replica pairs are going away in 1.6, but is master/slave >> still supposed to work? I can't get a slave to replicate using 1.5.6 >> as master. It starts up and begins to replicate, but seems to get >> stuck on the first DB it hits (which is very small). If I switch to >> using 1.4.4 as master, a 1.5.6 slave replicates fine.
>> The logs show it creating a bunch of temporary collections from the >> map-reduces we run, and I just see a bunch of this:
>> [replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was >> empty, skipping ahead local.sources >> [replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425 >> operations >> [replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38 >> 2010 4c4f4da6:6
>> Any suggestions?
>> Thanks, >> Ian
>> On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
>> > MongoDB 1.5.6 is out and ready for testing.
>> > We're wrapping up development for 1.6.0 and could use all the help we >> > can get testing this release. >> > Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x >> > version for Tuesday or Wednesday next week, and release 1.6.0 either >> > Friday 7/30 or Monday 8/2.
>> > Major features to test: >> > - Replica Sets : ready for testing. some loose ends we're still >> > fixing, but testers would be greatly appreciate >> > - Sharding : lots of stability/maturity improvements - testers would be great
> -- > You received this message because you are subscribed to the Google Groups "mongodb-user" group. > To post to this group, send email to mongodb-user@googlegroups.com. > To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
> The change we made is that the slave does a long poll to the master so > changes should be more real time rather than less.
> On Tue, Jul 27, 2010 at 11:28 PM, iragsdale <ian.ragsd...@gmail.com> wrote: >> Seems like master/slave still works, but the behavior is different. I >> reverted the master to 1.4.4 and started up the slave on 1.5.6. It >> did a full resync of around 100GB in a few hours. I then upgraded the >> master back to 1.5.6, and the slave seems to still be replicating, but >> instead of staying within a second or two of the master, it waits a >> full minute before grabbing the most recent changes, so it could be up >> to a minute behind at any point. Is this expected behavior?
>> - Ian
>> On Jul 27, 4:21 pm, iragsdale <ian.ragsd...@gmail.com> wrote: >>> I see that replica pairs are going away in 1.6, but is master/slave >>> still supposed to work? I can't get a slave to replicate using 1.5.6 >>> as master. It starts up and begins to replicate, but seems to get >>> stuck on the first DB it hits (which is very small). If I switch to >>> using 1.4.4 as master, a 1.5.6 slave replicates fine.
>>> The logs show it creating a bunch of temporary collections from the >>> map-reduces we run, and I just see a bunch of this:
>>> [replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was >>> empty, skipping ahead local.sources >>> [replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425 >>> operations >>> [replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38 >>> 2010 4c4f4da6:6
>>> Any suggestions?
>>> Thanks, >>> Ian
>>> On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
>>>> MongoDB 1.5.6 is out and ready for testing.
>>>> We're wrapping up development for 1.6.0 and could use all the help we >>>> can get testing this release. >>>> Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x >>>> version for Tuesday or Wednesday next week, and release 1.6.0 either >>>> Friday 7/30 or Monday 8/2.
>>>> Major features to test: >>>> - Replica Sets : ready for testing. some loose ends we're still >>>> fixing, but testers would be greatly appreciate >>>> - Sharding : lots of stability/maturity improvements - testers would be great
>> -- >> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >> To post to this group, send email to mongodb-user@googlegroups.com. >> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
> -- > You received this message because you are subscribed to the Google Groups "mongodb-user" group. > To post to this group, send email to mongodb-user@googlegroups.com. > To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
Its kind of wrong. That records the last time a sync pass finished. Since a sync pass now blocks for up to a minute in the server, that time could be anywhere from 0 to 60 seconds. Maybe we should add # of ops applied last time - that way you'll be able to see if it steady state or not.
> On Jul 27, 2010, at 10:31 PM, Eliot Horowitz wrote:
>> How are you measuring?
>> The change we made is that the slave does a long poll to the master so >> changes should be more real time rather than less.
>> On Tue, Jul 27, 2010 at 11:28 PM, iragsdale <ian.ragsd...@gmail.com> wrote: >>> Seems like master/slave still works, but the behavior is different. I >>> reverted the master to 1.4.4 and started up the slave on 1.5.6. It >>> did a full resync of around 100GB in a few hours. I then upgraded the >>> master back to 1.5.6, and the slave seems to still be replicating, but >>> instead of staying within a second or two of the master, it waits a >>> full minute before grabbing the most recent changes, so it could be up >>> to a minute behind at any point. Is this expected behavior?
>>> - Ian
>>> On Jul 27, 4:21 pm, iragsdale <ian.ragsd...@gmail.com> wrote: >>>> I see that replica pairs are going away in 1.6, but is master/slave >>>> still supposed to work? I can't get a slave to replicate using 1.5.6 >>>> as master. It starts up and begins to replicate, but seems to get >>>> stuck on the first DB it hits (which is very small). If I switch to >>>> using 1.4.4 as master, a 1.5.6 slave replicates fine.
>>>> The logs show it creating a bunch of temporary collections from the >>>> map-reduces we run, and I just see a bunch of this:
>>>> [replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was >>>> empty, skipping ahead local.sources >>>> [replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425 >>>> operations >>>> [replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38 >>>> 2010 4c4f4da6:6
>>>> Any suggestions?
>>>> Thanks, >>>> Ian
>>>> On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
>>>>> MongoDB 1.5.6 is out and ready for testing.
>>>>> We're wrapping up development for 1.6.0 and could use all the help we >>>>> can get testing this release. >>>>> Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x >>>>> version for Tuesday or Wednesday next week, and release 1.6.0 either >>>>> Friday 7/30 or Monday 8/2.
>>>>> Major features to test: >>>>> - Replica Sets : ready for testing. some loose ends we're still >>>>> fixing, but testers would be greatly appreciate >>>>> - Sharding : lots of stability/maturity improvements - testers would be great
>>>>> Thanks for the help and support, >>>>> -Eliot
>>> -- >>> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >>> To post to this group, send email to mongodb-user@googlegroups.com. >>> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >>> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
>> -- >> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >> To post to this group, send email to mongodb-user@googlegroups.com. >> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
> -- > You received this message because you are subscribed to the Google Groups "mongodb-user" group. > To post to this group, send email to mongodb-user@googlegroups.com. > To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
I've confirmed this by looking for new documents, and they do show up even when replication appears to be behind. So, I think we're good, the output of db.printSlaveReplicationInfo() is just a bit confusing.
However, that still doesn't explain why I couldn't get a full resync to work with 1.5.6 as the master, unless the query blocks for 60 seconds even when behind. In that case, it would probably take a really really long time to get a resync to happen, right?
- Ian
On Jul 27, 2010, at 11:16 PM, Eliot Horowitz wrote:
> Its kind of wrong. > That records the last time a sync pass finished. > Since a sync pass now blocks for up to a minute in the server, that > time could be anywhere from 0 to 60 seconds. > Maybe we should add # of ops applied last time - that way you'll be > able to see if it steady state or not.
> On Tue, Jul 27, 2010 at 11:33 PM, Ian Ragsdale <ian.ragsd...@gmail.com> wrote: >> I'm just looking at db.printSlaveReplicationInfo() - is that wrong?
>> On Jul 27, 2010, at 10:31 PM, Eliot Horowitz wrote:
>>> How are you measuring?
>>> The change we made is that the slave does a long poll to the master so >>> changes should be more real time rather than less.
>>> On Tue, Jul 27, 2010 at 11:28 PM, iragsdale <ian.ragsd...@gmail.com> wrote: >>>> Seems like master/slave still works, but the behavior is different. I >>>> reverted the master to 1.4.4 and started up the slave on 1.5.6. It >>>> did a full resync of around 100GB in a few hours. I then upgraded the >>>> master back to 1.5.6, and the slave seems to still be replicating, but >>>> instead of staying within a second or two of the master, it waits a >>>> full minute before grabbing the most recent changes, so it could be up >>>> to a minute behind at any point. Is this expected behavior?
>>>> - Ian
>>>> On Jul 27, 4:21 pm, iragsdale <ian.ragsd...@gmail.com> wrote: >>>>> I see that replica pairs are going away in 1.6, but is master/slave >>>>> still supposed to work? I can't get a slave to replicate using 1.5.6 >>>>> as master. It starts up and begins to replicate, but seems to get >>>>> stuck on the first DB it hits (which is very small). If I switch to >>>>> using 1.4.4 as master, a 1.5.6 slave replicates fine.
>>>>> The logs show it creating a bunch of temporary collections from the >>>>> map-reduces we run, and I just see a bunch of this:
>>>>> [replslave] Tue Jul 27 21:20:09 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:10 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:12 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:13 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:14 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:15 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:20 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:26 info DFM::findAll(): extent 0:2700 was >>>>> empty, skipping ahead local.sources >>>>> [replslave] Tue Jul 27 21:20:38 repl: checkpoint applied 1425 >>>>> operations >>>>> [replslave] Tue Jul 27 21:20:38 repl: syncedTo: Tue Jul 27 21:20:38 >>>>> 2010 4c4f4da6:6
>>>>> Any suggestions?
>>>>> Thanks, >>>>> Ian
>>>>> On Jul 24, 12:29 pm, Eliot Horowitz <eliothorow...@gmail.com> wrote:
>>>>>> MongoDB 1.5.6 is out and ready for testing.
>>>>>> We're wrapping up development for 1.6.0 and could use all the help we >>>>>> can get testing this release. >>>>>> Depending on what we find, we'll be targeting 1.5.7 as the final 1.5.x >>>>>> version for Tuesday or Wednesday next week, and release 1.6.0 either >>>>>> Friday 7/30 or Monday 8/2.
>>>>>> Major features to test: >>>>>> - Replica Sets : ready for testing. some loose ends we're still >>>>>> fixing, but testers would be greatly appreciate >>>>>> - Sharding : lots of stability/maturity improvements - testers would be great
>>>>>> Thanks for the help and support, >>>>>> -Eliot
>>>> -- >>>> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >>>> To post to this group, send email to mongodb-user@googlegroups.com. >>>> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >>>> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
>>> -- >>> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >>> To post to this group, send email to mongodb-user@googlegroups.com. >>> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >>> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
>> -- >> You received this message because you are subscribed to the Google Groups "mongodb-user" group. >> To post to this group, send email to mongodb-user@googlegroups.com. >> To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. >> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
> -- > You received this message because you are subscribed to the Google Groups "mongodb-user" group. > To post to this group, send email to mongodb-user@googlegroups.com. > To unsubscribe from this group, send email to mongodb-user+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.
