RE: [PHP-DB] database connection timeout
| From: | Juffermans, Jos | Date: | Thu, 24 Mar 2005 15:06:23 +0000 |
| Subject: | RE: [PHP-DB] database connection timeout | ||
| Groups: | php.db | ||
| Request: | Send a blank email to php-db+get-39024@lists.php.net to get a copy of this message | ||
Hi,
Martin Norland wrote:
MN> I'm afraid phrasing a question multiple ways over a course of days tends
MN> not to have much success. In an effort to avoid populating google with
MN> just the question, I will give to you what I would try.
I understand your point but since I wasn't getting any response and never
received a confirmation email from the system telling me that my email
account had been verified, I wasn't sure that the original post was actually
sent to the community. Oddly enough I got an email from the system this
morning asking me to verify my email address - again :s
Anyway, at least this time it triggered a response ;)
MN> Is it that it's not always available, or that it needs time to 'come
MN> back up'. You may want to try running some random low-traffic
MN> application over it to make sure it remains running. Some simple SNMP
MN> protocol or even simpler just connect to a remote GKrellM. Heck, a ssh
MN> connection with top running would do - just something that is constantly
MN> transmitting a trickle of data.
The VPN connection is not always available. I'm afraid I don't have much
control over the VPN myself and I believe only the Oracle port has been
opened in this VPN.
MN> Essentially, you want ocilogon (complex network function to login to a
MN> networked Oracle [in this case] server) to timeout just like fsockopen
MN> (fairly simple function to open a socket, including network resources).
MN> My suggestion is to try connecting using fsockopen, with the timeout.
MN> If it successfully connects - chances are, the majority of the time,
MN> your ocilogon that follows will work as well.
That's a good idea, I'll try to connect to the Oracle port thru fsockopen.
MN> Another dirty little trick you might use is running the ocilogon in a
MN> separate script and making use of set_time_limit() to make sure it
MN> doesn't run too long. Just be sure not to set_time_limit() for your
MN> whole set of scripts, unless you really want them to abort after the 5
MN> or 10 second limit you're setting ocilogon (unlikely).
I'm not sure if I can set the time limit (safemode...) but are you saying
that set_time_limit only affects the script that you include and not the
caller script?
I'll try the fsockopen "connection test" on the oracle port. Thanks for your
help.
Rgds,
Jos