From: "Bob Beilstein" Subject: Re: CHECKPOINTS Date: 2000/03/03 Message-ID: #1/1 References: <89k4d5$di0$1@news.online.de> <38bdee81.39787591@news.eagles.bbs.net.au> X-Priority: 3 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600 X-Complaints-To: abuse@earthlink.net X-Trace: newsread2.prod.itd.earthlink.net 952052517 168.191.59.158 (Thu, 02 Mar 2000 19:01:57 PST) Organization: EarthLink Inc. -- http://www.EarthLink.net X-MSMail-Priority: Normal NNTP-Posting-Date: Thu, 02 Mar 2000 19:01:57 PST Newsgroups: comp.databases.oracle.server Hi, Steve Thank you for pointing out the fact that ENABLE_CHECKPOINT_PROCESS will not "fix" Mario's problem. I agree. I should have made it more clear that this suggestion was just something in addition to more substantive action to improve I/O performance. I was well aware that the CKPT process only frees up the LGWR somewhat, and that Mario's problem probably lies in the DBWR I/O realm. On the other hand, enabling the CKPT process does, in my experience, help some (which is, I assume, why Oracle 8 made it a standard feature rather than an option). On occasion, "some" is enough to push the problem off (on rare occasion, enough to actually "cure" it). Craig's paper includes information about I/O and so forth -- so I was just adding a couple of cents worth about something in the "can't hurt, might help" category. I certainly would not expect turning on ENABLE_CHECKPOINT_PROCESS to have any dramatic effect on its own. Thanks again. / Bob "Steve Adams" wrote in message news:38bdee81.39787591@news.eagles.bbs.net.au... > Hi Bob, > > The checkpoint process is only responsible for the writes to the > datafile headers, not the data block writes. > Enabling it would not help Mario at all. > > Regards, > Steve Adams > http://www.ixora.com.au/ > http://www.oreilly.com/catalog/orinternals/ > http://www.christianity.net.au/ > > > On Thu, 02 Mar 2000 03:57:10 GMT, "Bob Beilstein" > wrote: > > >In addition to Craig Shallhamer's excellent suggestions, you want to make > >sure that you have: > > > >CHECKPOINT_PROCESS=TRUE > > > >set in your init.ora. This will start a separate process devoted to doing > >checkpoint I/O. > > > >In my experience, it can make the difference between getting the log files > >emptied before you roll back to them, and not. > > > >You also didn't mention how big your logfiles are. We've found that for a > >lot of databases, increasing the size of the logfiles (we often use 64MB or > >larger for big databases) can also help. Your mileage may vary. > > > > > > > >"Mario" wrote in message > >news:89k4d5$di0$1@news.online.de... > >> > >> We have a problem with redo log switching > >> when running large jobs. All groups fill up > >> and remain active even though they > >> are archived - presumably this is because > >> checkpointing is going on. > >> > >> Finally when the > >> last log group fills and wants to switch back to > >> the first group it cannot do so, because > >> the first group is still active. At this point the system > >> slows to a crawl and even trivial queries require minutes > >> for a response. > >> > >> We have > >> increased the number of redo log groups > >> (almost doubling the space available), > >> and have also increased the initialization > >> parameter db_block_checkpoint_batch > >> so as to allow for faster checkpointing. > >> Yet the problem persists. > >> > >> Does anyone have any pointers for us? > >> > >> Is there any good documentation on how the checkpoint > >> process works and why it takes so long? > >> > >> Our System > >> Oracle V7.3.4, RS/6000, 2CPUs, 1.3 GB Memory, AIX > >> > >> Thanks in advance, > >> Mario > >> > >> > >> > >> > >> > > > > >