Thread: postmaster: StreamConnection: accept: No such device or address
Using a legacy installation ( 7.2.3 ). Occasionally the system will reach a state where attempted psql connection attempts fail, with the following error in the postgresql log: postmaster: StreamConnection: accept: No such device or address Will also occasionally get "no connection to server" errors from ecpg programs ( which I believe are probably due to the same issud ). Can anyone shed some light on what could be causing this? OS resource limitation? Known bug in ancient version of PG? thanks, reid
Reid Thompson wrote: > Using a legacy installation ( 7.2.3 ). > Occasionally the system will reach a state where attempted psql > connection attempts fail, with the following error in the postgresql > log: > postmaster: StreamConnection: accept: No such device or address > > Will also occasionally get "no connection to server" errors from ecpg > programs ( which I believe are probably due to the same issud ). > > Can anyone shed some light on what could be causing this? OS resource > limitation? Known bug in ancient version of PG? > > thanks, > reid Is something cleaning /tmp of the unix socket? Older versions didn't keep the timestamp updated, and /tmp cleaning utilities could remove the socket. -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 512-248-2683 E-Mail: ler@lerctr.org US Mail: 430 Valona Loop, Round Rock, TX 78681-3893
Larry Rosenman wrote: > Reid Thompson wrote: > >> Using a legacy installation ( 7.2.3 ). >> Occasionally the system will reach a state where attempted psql >> connection attempts fail, with the following error in the postgresql >> log: >> postmaster: StreamConnection: accept: No such device or address >> >> Will also occasionally get "no connection to server" errors from ecpg >> programs ( which I believe are probably due to the same issud ). >> >> Can anyone shed some light on what could be causing this? OS resource >> limitation? Known bug in ancient version of PG? >> >> thanks, >> reid >> > > > Is something cleaning /tmp of the unix socket? Older versions didn't keep > the > timestamp updated, and /tmp cleaning utilities could remove the socket. > > > The errors occur when psql is used with -h also, and the ecpg program is connecting from a different host.