DBServer name not found -- SOMETIMES

Hi All, We recently recovered from a nasty controller failure, leaving some of the data corrupt. Though, luckily, none of it was database related (checked w/ e2fsck AND oncheck). This is Online version 7.31.Something.Or.Other. We're back in business, or so I thought, with a dual Athlon machine running a stock Redhat 2.4.28 kernel, from Fedora Legacy. We also run PostgreSQL and MySQL on the same box, but for the moment, I've disabled them from loading. When I start up informix with an oninit -v, it fails on forking the 1 ipcshm listener, with a log of: 10:51:49 Event alarms enabled. ALARMPROG = '/opt/informix/etc/eventalarm' 10:51:57 DR: DRAUTO is 0 (Off) 10:51:57 listener-thread: err = -25555: oserr = 0: errstr = london:Server london is not listed as a dbserver name in sqlhosts. 10:51:57 Attempting to bring listener thread down. 10:51:57 Server stopped. This is concerning, b/c my SQL Hosts file looks like: london onipcshm corp london london_net onsoctcp corp london_net And obviously, we DO have a london server in the sqlhosts file. It's even finding that file, according to strace: read(0, "\\4\\0\\0\\0/opt/informix/etc/onconfig.a"..., 136) = 136 read(0, "\\5\\0\\0\\0corp.inventconnect.com\\0\\0\\0\\0\\0\\0"..., 136) = 136 read(0, "\\6\\0\\0\\0Informix Dynamic Server Vers"..., 136) = 136 read(0, "", 136) = 0 lseek(0, 0, SEEK_END) = 680 write(0, "\\10\\0\\0\\0/opt/informix/etc/sqlhosts\\0s"..., 136) = 136 I also have an entry for london_net 1314/tcp in /etc/services. I tried changing corp to the IP address, no luck. But, this doesn't happen ALL the time. It happened once, last week, and we bounced the server [for a different] reason, and it came up w/ no issues. This time, however, I can't get it to come up at all (and we tried bouncing it). Any thoughts? --Anthony

I realize it's bad kharma to reply to your own post, but I have some more interesting pieces to the puzzle. I have a VMWare instance that I upgraded to the same kernel, and it runs fine. It's also running the same version of informix, which appears fine. However, when stepping through a Strace, informix does NOT read more than the first valid line in the sqlhosts file. It doesn't make sense to me, but maybe someone else can help out. --Anthony anthony@zoraptera.com (Anthony Presley) wrote in message news:<a55ad738.0410220720.5f76dae0@posting.google.com>... > Hi All, > > We recently recovered from a nasty controller failure, leaving some of > the data corrupt. Though, luckily, none of it was database related > (checked w/ e2fsck AND oncheck). This is Online version > 7.31.Something.Or.Other. > > We're back in business, or so I thought, with a dual Athlon machine > running a stock Redhat 2.4.28 kernel, from Fedora Legacy. We also run > PostgreSQL and MySQL on the same box, but for the moment, I've > disabled them from loading. > > When I start up informix with an oninit -v, it fails on forking the 1 > ipcshm listener, with a log of: > > 10:51:49 Event alarms enabled. ALARMPROG = > '/opt/informix/etc/eventalarm' > 10:51:57 DR: DRAUTO is 0 (Off) > 10:51:57 listener-thread: err = -25555: oserr = 0: errstr = london:> Server london is not listed as a dbserver name in sqlhosts. > > 10:51:57 Attempting to bring listener thread down. > > 10:51:57 Server stopped.> > This is concerning, b/c my SQL Hosts file looks like: > > london onipcshm corp london > london_net onsoctcp corp london_net> > And obviously, we DO have a london server in the sqlhosts file. It's > even finding that file, according to strace: > > read(0, "\\4\\0\\0\\0/opt/informix/etc/onconfig.a"..., 136) = 136 > read(0, "\\5\\0\\0\\0corp.inventconnect.com\\0\\0\\0\\0\\0\\0"..., 136) = 136 > read(0, "\\6\\0\\0\\0Informix Dynamic Server Vers"..., 136) = 136 > read(0, "", 136) = 0 > lseek(0, 0, SEEK_END) = 680 > write(0, "\\10\\0\\0\\0/opt/informix/etc/sqlhosts\\0s"..., 136) = 136 > > I also have an entry for london_net 1314/tcp in /etc/services. > > I tried changing corp to the IP address, no luck. > > But, this doesn't happen ALL the time. It happened once, last week, > and we bounced the server [for a different] reason, and it came up w/ > no issues. This time, however, I can't get it to come up at all (and > we tried bouncing it). > > Any thoughts? > > --Anthony

First thing to try after reading your 2nd posting is : set list in your sqlhosts file and look for any kind of special characters. or just create a new one. anthony@zoraptera.com (Anthony Presley) wrote in message news:<a55ad738.0410221521.771a4097@posting.google.com>... > I realize it's bad kharma to reply to your own post, but I have some > more interesting pieces to the puzzle. > > I have a VMWare instance that I upgraded to the same kernel, and it > runs fine. It's also running the same version of informix, which > appears fine. > > However, when stepping through a Strace, informix does NOT read more > than the first valid line in the sqlhosts file. It doesn't make sense > to me, but maybe someone else can help out. > > --Anthony > > anthony@zoraptera.com (Anthony Presley) wrote in message news:<a55ad738.0410220720.5f76dae0@posting.google.com>... > > Hi All, > > > > We recently recovered from a nasty controller failure, leaving some of > > the data corrupt. Though, luckily, none of it was database related > > (checked w/ e2fsck AND oncheck). This is Online version > > 7.31.Something.Or.Other. > > > > We're back in business, or so I thought, with a dual Athlon machine > > running a stock Redhat 2.4.28 kernel, from Fedora Legacy. We also run > > PostgreSQL and MySQL on the same box, but for the moment, I've > > disabled them from loading. > > > > When I start up informix with an oninit -v, it fails on forking the 1 > > ipcshm listener, with a log of: > > > > 10:51:49 Event alarms enabled. ALARMPROG = > > '/opt/informix/etc/eventalarm' > > 10:51:57 DR: DRAUTO is 0 (Off) > > 10:51:57 listener-thread: err = -25555: oserr = 0: errstr = london:> > Server london is not listed as a dbserver name in sqlhosts. > > > > 10:51:57 Attempting to bring listener thread down. > > > > 10:51:57 Server stopped.> > > > This is concerning, b/c my SQL Hosts file looks like: > > > > london onipcshm corp london > > london_net onsoctcp corp london_net> > > > And obviously, we DO have a london server in the sqlhosts file. It's > > even finding that file, according to strace: > > > > read(0, "\\4\\0\\0\\0/opt/informix/etc/onconfig.a"..., 136) = 136 > > read(0, "\\5\\0\\0\\0corp.inventconnect.com\\0\\0\\0\\0\\0\\0"..., 136) = 136 > > read(0, "\\6\\0\\0\\0Informix Dynamic Server Vers"..., 136) = 136 > > read(0, "", 136) = 0 > > lseek(0, 0, SEEK_END) = 680 > > write(0, "\\10\\0\\0\\0/opt/informix/etc/sqlhosts\\0s"..., 136) = 136 > > > > I also have an entry for london_net 1314/tcp in /etc/services. > > > > I tried changing corp to the IP address, no luck. > > > > But, this doesn't happen ALL the time. It happened once, last week, > > and we bounced the server [for a different] reason, and it came up w/ > > no issues. This time, however, I can't get it to come up at all (and > > we tried bouncing it). > > > > Any thoughts? > > > > --Anthony