informix performance degrade after one of out EMC drive fail

Hi all, Our scripts that use informix server as the DBMS become very slow after one of our EMC hard disk fail. We have not change the failed disk yet because currently a spare EMC disk is temporarily replacing it. Some of our jobs that normally took 4 hours to run now seems to can only finished after 10 hours. We have perform update statistics on the tables being used by the scripts but it does not help much. Can anyone please help us? This server is a production server and we cannot afford our jobs to be continue suspended like that. Thanks a lot! ps: we have tried to reboot the informix server as well as the host server but it does not help much ps: our dba claim that the failure of EMC disk will not cause performance degradation Thanks

come on..... after a disk failure on emc the whole world comes down... you could start looking at sar/iostat which will probably tell you that your i/o subsystem is the slow factor. Superboer. ggk517@gmail.com schreef: > Hi all, > > Our scripts that use informix server as the DBMS become very slow after > one of our EMC hard disk fail. We have not change the failed disk yet > because currently a spare EMC disk is temporarily replacing it. Some of > our jobs that normally took 4 hours to run now seems to can only > finished after 10 hours. We have perform update statistics on the > tables being used by the scripts but it does not help much. Can anyone > please help us? This server is a production server and we cannot afford > our jobs to be continue suspended like that. Thanks a lot! > > ps: we have tried to reboot the informix server as well as the host > server but it does not help much > > ps: our dba claim that the failure of EMC disk will not cause > performance degradation > > Thanks

<ggk517@gmail.com> wrote in message news:1152104152.291953.206670@j8g2000cwa.googlegroups.com... > Hi all, > > Our scripts that use informix server as the DBMS become very slow after > one of our EMC hard disk fail. We have not change the failed disk yet > because currently a spare EMC disk is temporarily replacing it. Some of > our jobs that normally took 4 hours to run now seems to can only > finished after 10 hours. We have perform update statistics on the > tables being used by the scripts but it does not help much. Can anyone > please help us? This server is a production server and we cannot afford > our jobs to be continue suspended like that. Thanks a lot! > > ps: we have tried to reboot the informix server as well as the host > server but it does not help much > > ps: our dba claim that the failure of EMC disk will not cause > performance degradation How do you know that you have an EMC disk failure? Presumably you have one physical disk in a Standard EMC pair, that contains one or more hypervolumes/LUNs presented to your host, that has failed? In that case, I'd be inclined to agree with your DBA.

Neil Truby wrote: > <ggk517@gmail.com> wrote in message > news:1152104152.291953.206670@j8g2000cwa.googlegroups.com... >> Hi all, >> >> Our scripts that use informix server as the DBMS become very slow after >> one of our EMC hard disk fail. We have not change the failed disk yet >> because currently a spare EMC disk is temporarily replacing it. Some of >> our jobs that normally took 4 hours to run now seems to can only >> finished after 10 hours. We have perform update statistics on the >> tables being used by the scripts but it does not help much. Can anyone >> please help us? This server is a production server and we cannot afford >> our jobs to be continue suspended like that. Thanks a lot! >> >> ps: we have tried to reboot the informix server as well as the host >> server but it does not help much >> >> ps: our dba claim that the failure of EMC disk will not cause >> performance degradation > > How do you know that you have an EMC disk failure? Presumably you have one > physical disk in a Standard EMC pair, that contains one or more > hypervolumes/LUNs presented to your host, that has failed? > In that case, I'd be inclined to agree with your DBA. > > What does the System Admin geezer say :p ? What does dd say? What actually "broke"? "What changed?" "Nothing!" "Nothing?" "Well, the disks broke a bit; but it can't be related to that!" What type of set up do these disks have??? RAID 10, RAID 5?? Nothing?? Is there a "cache" here??

ggk517@gmail.com wrote: > Hi all, > > Our scripts that use informix server as the DBMS become very slow after > one of our EMC hard disk fail. We have not change the failed disk yet > because currently a spare EMC disk is temporarily replacing it. Some of > our jobs that normally took 4 hours to run now seems to can only > finished after 10 hours. We have perform update statistics on the > tables being used by the scripts but it does not help much. Can anyone > please help us? This server is a production server and we cannot afford > our jobs to be continue suspended like that. Thanks a lot! > > ps: we have tried to reboot the informix server as well as the host > server but it does not help much > > ps: our dba claim that the failure of EMC disk will not cause > performance degradation Your DBA is incorrect, but the performance hit depends on the type of array that's been configured on the EMC. If it's RAID10 there should be little if any noticeable effect on performance except under extreme load. If, however, you've been coerced into using RAID5 performance will degrade by at least 20-25% until you replace the drive and as much as 80% during the replacement drive rebuild. NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! NO RAID5!! Art S. Kagel