Re: bind 9 performance
Delegation records caught us too. There used to be a document called something like "top 10 things to pay attention to when you upgrade from bind 8 to bind 9" which included this gotcha, and I'd wished I'd paid real attention to it. But it was easily fixed once the problem was understood. We found bind 9 to use more memory and cpu, but this was easily handled by normal server replacement. If your site has not too much of anything specific, e.g. 20 zones, 2000 names, 500 desktops using it, one public website with 10,000 hits a day, I expect it is extremely difficult to buy a current computer slow enough to have any trouble with bind 9. A modern $300 laptop would be practically idle. But if your server is very old, or if there is some aspect of your configuration/site that is scaled up or advanced features that you use, you may have to read more, tune, benchmark, etc. If your spam filter retrieves its data via dns records, that could push up your query rate and cache size. John Wobus On Jun 15, 2011, at 5:52 PM, Mark K. Pettit wrote: One of the things that got us is we didn't know BIND 8 automatically created delegation records in a zone at the zone cut, if the nameserver knew of the existence of the cut. For example, if we have the following zones in our named.conf: zone "example.com" { ... }; zone "sub.example.com" { ... }; and inside example.com, we do *not* have any delegation records for "sub.example.com", BIND 8 will automatically create the NS records in example.com. BIND 9 will not do this. Be sure all of your zones that have zone cuts also have the proper delegation records. On Jun 15, 2011, at 1:30 PM, Eivind Olsen wrote: abushla...@ies.etisalat.ae wrote: What about zone configuration in BIND 8 and BIND 9? Is there any difference between the two ? Do you mean the zone configuration in named.conf, or the zonefiles? BIND9 has a doc/misc/migration document which gives plenty of good advice on configuration changes from BIND8 to BIND9. In general, what I'd recommend is: 1) Read that migration document 2) Test your existing named.conf + zonefiles by either loading them into BIND9 directly, or by using the named-checkconf / named-checkzone commands from BIND9. 3) Watch your logs Regards Eivind Olsen ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users ___ Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
Re: bind 9 performance
One of the things that got us is we didn't know BIND 8 automatically created delegation records in a zone at the zone cut, if the nameserver knew of the existence of the cut. For example, if we have the following zones in our named.conf: zone "example.com" { ... }; zone "sub.example.com" { ... }; and inside example.com, we do *not* have any delegation records for "sub.example.com", BIND 8 will automatically create the NS records in example.com. BIND 9 will not do this. Be sure all of your zones that have zone cuts also have the proper delegation records. On Jun 15, 2011, at 1:30 PM, Eivind Olsen wrote: > abushla...@ies.etisalat.ae wrote: > >> What about zone configuration in BIND 8 and BIND 9? Is there any >> difference between the two ? > > Do you mean the zone configuration in named.conf, or the zonefiles? > > BIND9 has a doc/misc/migration document which gives plenty of good advice > on configuration changes from BIND8 to BIND9. > > In general, what I'd recommend is: > > 1) Read that migration document > 2) Test your existing named.conf + zonefiles by either loading them into > BIND9 directly, or by using the named-checkconf / named-checkzone commands > from BIND9. > 3) Watch your logs > > Regards > Eivind Olsen > > > ___ > bind-users mailing list > bind-users@lists.isc.org > https://lists.isc.org/mailman/listinfo/bind-users ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
RE: bind 9 performance
hugo hugoo wrote: > Is there anything I have to look at to check that all is OK in terms of > performances when I will be in BIND9? Well, you haven't really given any information about your current setup and usage, so I have no idea if you're trying to run a million-user ISP with DNS on an old Intel 486 processor with 8MB RAM :-) In general, if you want good performance, generally without really having to tune your operating system etc, the easy answer is: make sure BIND has enough physical memory available for its cache, and run it on a system with 4 CPU cores available (BIND9 doesn't really seem to scale much past 4 cores). If you're really concerned about performance: run benchmarks. Two interesting things to measure are number of queries per second, and response time/latency. Keep in mind what you're measuring as well - if you query from the same server and ask for information avaialble in the cache, you're just doing a local test. A real life query might very well make your nameserver ask several external servers, which takes time no matter what you do. External factors can and will impact your benchmarking. Know what you're actually benchmarking. One example: I was doing some benchmarking a while ago, and was confused as to why I could only get something like 3500 responses per second from a server when I asked it to give some >3Kbyte answers back - which was actually a very nice number, considering I did the testing on a 100Mbps network link and managed to fill it to just over 97% of maximum theoretical limit. Please provide more information if you want a more accurate answer. 1) Which role does your BIND server have? Authoritative server only? Recursive server only? Mix of both? 2) Is it a plain setup, or do you do anything fancy, like feeding BIND from a database backend? 3) Please describe the server you're running BIND on. Is it single-core CPU or multi-core? How much physical memory does it have? Does it basically just run BIND, or does BIND have to compete with "a ton" of other services for resources? 4) What kind of use does your nameserver see? Number of users? Queries per second? If you don't know - dump statistics during prime time for your users and check - you should ideally be able to extract some statistics from BIND8 by running "ndc stats" and then looking into named.stats wherever your nameserver has its working directory. If you found the file - wait a while, run the command again to dump another set of statistics - you should now have some numbers available in that file and can now calculate how many queries you have per second on average. Regards Eivind Olsen ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
RE: bind 9 performance
I do not change the zone files. I only remove some logging category nt compatible with bind9, that's all. I agree that I have to go to BIND9. My question was related to the fact that I am a little worry about a difference in performance when I will be in BIND9. So I wonder if I do not have to also upgrade the hardware. Is there anything I have to look at to check that all is OK in terms of performances when I will be in BIND9? Regards, Hugo, > Subject: Re: bind 9 performance > To: eiv...@aminor.no; > bind-users-bounces+abushlaibi=ies.etisalat...@lists.isc.org; > bind-users@lists.isc.org > From: abushla...@ies.etisalat.ae > Date: Wed, 15 Jun 2011 20:06:11 + > > What about zone configuration in BIND 8 and BIND 9? Is there any difference > between the two ? > > Thanks & Regards > > -Original Message- > From: Eivind Olsen > Sender: bind-users-bounces+abushlaibi=ies.etisalat...@lists.isc.org > Date: Wed, 15 Jun 2011 20:30:58 > To: > Subject: Re: bind 9 performance > > hugo hugoo wrote: > > > - Has anyone faced a performance problem due to an upgrade bind8/bind9? > > I didn't notice anything like that when I last upgraded from BIND8 (back > in 2001 or so). > > When that is said: what kind of hardware are you running it on? Single > CPU? Multiple cores? I've seen some fairly ancient performance comparisons > between BIND8 and BIND9 that claimed BIND8 performed best on a single-core > CPU only. And if I'm not mistaken, BIND9 had some less-than-optimal > threading back in those days, and performs much better now. > > Is pure performance important to you? Will you be likely to even notice? > Also, BIND8 has been deprecated for almost 4 years now. > > Regards > Eivind Olsen > > > ___ > bind-users mailing list > bind-users@lists.isc.org > https://lists.isc.org/mailman/listinfo/bind-users > ___ > bind-users mailing list > bind-users@lists.isc.org > https://lists.isc.org/mailman/listinfo/bind-users ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
Re: bind 9 performance
abushla...@ies.etisalat.ae wrote: > What about zone configuration in BIND 8 and BIND 9? Is there any > difference between the two ? Do you mean the zone configuration in named.conf, or the zonefiles? BIND9 has a doc/misc/migration document which gives plenty of good advice on configuration changes from BIND8 to BIND9. In general, what I'd recommend is: 1) Read that migration document 2) Test your existing named.conf + zonefiles by either loading them into BIND9 directly, or by using the named-checkconf / named-checkzone commands from BIND9. 3) Watch your logs Regards Eivind Olsen ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
Re: bind 9 performance
On 6/15/2011 4:06 PM, abushla...@ies.etisalat.ae wrote: What about zone configuration in BIND 8 and BIND 9? Is there any difference between the two ? Thanks& Regards -Original Message- From: Eivind Olsen Sender: bind-users-bounces+abushlaibi=ies.etisalat...@lists.isc.org Date: Wed, 15 Jun 2011 20:30:58 To: Subject: Re: bind 9 performance hugo hugoo wrote: - Has anyone faced a performance problem due to an upgrade bind8/bind9? I didn't notice anything like that when I last upgraded from BIND8 (back in 2001 or so). When that is said: what kind of hardware are you running it on? Single CPU? Multiple cores? I've seen some fairly ancient performance comparisons between BIND8 and BIND9 that claimed BIND8 performed best on a single-core CPU only. And if I'm not mistaken, BIND9 had some less-than-optimal threading back in those days, and performs much better now. Is pure performance important to you? Will you be likely to even notice? Also, BIND8 has been deprecated for almost 4 years now. See doc/misc/migration. - Kevin ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
Re: bind 9 performance
What about zone configuration in BIND 8 and BIND 9? Is there any difference between the two ? Thanks & Regards -Original Message- From: Eivind Olsen Sender: bind-users-bounces+abushlaibi=ies.etisalat...@lists.isc.org Date: Wed, 15 Jun 2011 20:30:58 To: Subject: Re: bind 9 performance hugo hugoo wrote: > - Has anyone faced a performance problem due to an upgrade bind8/bind9? I didn't notice anything like that when I last upgraded from BIND8 (back in 2001 or so). When that is said: what kind of hardware are you running it on? Single CPU? Multiple cores? I've seen some fairly ancient performance comparisons between BIND8 and BIND9 that claimed BIND8 performed best on a single-core CPU only. And if I'm not mistaken, BIND9 had some less-than-optimal threading back in those days, and performs much better now. Is pure performance important to you? Will you be likely to even notice? Also, BIND8 has been deprecated for almost 4 years now. Regards Eivind Olsen ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
Re: bind 9 performance
hugo hugoo wrote: > - Has anyone faced a performance problem due to an upgrade bind8/bind9? I didn't notice anything like that when I last upgraded from BIND8 (back in 2001 or so). When that is said: what kind of hardware are you running it on? Single CPU? Multiple cores? I've seen some fairly ancient performance comparisons between BIND8 and BIND9 that claimed BIND8 performed best on a single-core CPU only. And if I'm not mistaken, BIND9 had some less-than-optimal threading back in those days, and performs much better now. Is pure performance important to you? Will you be likely to even notice? Also, BIND8 has been deprecated for almost 4 years now. Regards Eivind Olsen ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
bind 9 performance
Hello all, I plan to replace bind8 with bind9 on a same hardware (just software upgrade). - Has anyone faced a performance problem due to an upgrade bind8/bind9? - Is bind9 less performant or do I have to be confident on this aspect? Thanks in advance for any feedback, Hugo, ___ bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users